Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author.

Biofilm formation by B. licheniformis isolated from

whey protein concentrate 80 powder as a potential

source of product contamination

A thesis presented in partial fulfilment of the requirements for the degree of

Doctor of Philosophy in Food Technology

at Massey University, Palmerston North New Zealand

Siti Norbaizura binti Md Zain

2018

ABSTRACT

This study aimed to examine biofilm formation of licheniformis isolated from whey protein concentrate 80 (WPC80) as a potential source of contamination in the manufacture of WPC.

Six WPC80 powder samples from one whey processing plant in New Zealand were used in this study. Six Bacillus species including (percentage of isolates in brackets) B. licheniformis (66%), Bacillus cereus/Bacillus thuringiensis (18%), Bacillus subtilis (4%), Bacillus pumilus (4%), Paenibacillus glucanolyticus (2%) and Lactobacillus plantarum (6%) were identified using colony morphologies, biochemical tests, species specific PCR and 16S ribosomal DNA gene sequencing and subsequent analysis using the BLAST and Seqmatch databases.

Preliminary screening for biofilm formation by the predominant contaminant, B. licheniformis using a microtitre plate assay with the grown in laboratory medium tryptic soy broth (TSB) at three different temperatures (30°C, 37°C and 55°C) showed most biofilm formation at 37°C with 9/33 isolates forming strong biofilm. In total 13/33 isolates formed strong biofilm at three different temperatures on the polystyrene microtitre plate surface.

Subsequent tests for biofilm formation on stainless steel (SS) showed an increased frequency of biofilm formation with 32/33 strains forming strong biofilm in TSB at 37°C. This demonstrates the limitation of the microtitre plate assay for screening for biofilm formation and suggests that biofilm growth of B. licheniformis favours a SS surface.

The attachment and biofilm formation was further investigated using SS coupons and reconstituted whey medium at different concentrations (1%, 5%, and 20%). The best medium for B. licheniformis isolates to form biofilm on SS at its best growth temperature (37°C) was 1% reconstituted WPC80. Interestingly, when 1% reconstituted WPC80 was supplemented with lactose and minerals (mainly calcium and magnesium)

1

to replicate the composition of Mozzarella cheese whey before ultrafiltration (UF), the B. licheniformis biofilm counts increased at least by one log.

The production of protease enzyme, extracellular polymeric substances (EPS) and nitrate reduction by B. licheniformis showed the potential of B. licheniformis to influence the quality of dairy products. Biosurfactant production by B. licheniformis identified as lichenysin consisting of lipopeptide was detected and this may influence biofilm formation on SS. The inability of the B. licheniformis isolates to ferment lactose as their major carbon source was confirmed by lactose fermentation tests and shows that B. licheniformis is not ideally suited to a dairy environment. The B. licheniformis vegetative cells were found to be heat resistant with a < log10 reduction at the three temperatures tested; 72oC, 75oC and 80°C during 15 s, 30 s and 60 s heating intervals.

In order to thrive in a dairy system, synergistic interactions with other microflora were investigated as a possible mechanism to use lactose that has been broken down by other microflora. Lactobacillus plantarum (L. plantarum), another isolate from the WPC80 samples, has the ability to produce and galactose from lactose. This was grown with each of two B. licheniformis isolates (E30C11 and F30C02) with different abilities to form biofilm. Interestingly this did not enhance the growth of B. licheniformis suggesting that another carbon source, most likely whey protein, must provide the energy source for this bacterium in a whey environment.

A review of the WPC80 processing plant showed the UF membranes had the largest surface area (3500 – 7500 m2), providing most potential for biofilm growth. However, UF was run at 10°C, too low for the growth of B. licheniformis which has a minimum growth temperature of 20°C. The hypothesis that sections of the processing plant before the UF step are the sites for B. licheniformis biofilm growth was supported by analysing several samples from the raw whey balance tank, clarifier, thermaliser and separator where 7 B. licheniformis strains were isolated. This shows that B. licheniformis is present at several early stages of WPC processing, with the most likely areas for growth being the certain sections of the clarifier, thermaliser and the separator where temperatures are close to the best growth temperature for this bacterium (37°C).

2

Preventing B. licheniformis contamination of WPC needs to focus on adjusting the conditions in these sections of the processing plant to limit biofilm growth.

Keywords: dairy, Bacillus species, L. plantarum, lichenysin, stainless steel, membrane processing plant.

3

FRONTISPIECE Biofilm of Bacillus licheniformis embedded in extracellular

polymeric substances on 304 grade stainless steel after 24 h

incubation.

4

LIST OF PUBLICATIONS

This work has been published in part in the following publications:

1. Xuemei Tang, Steve Flint, Rod Bennett, John Brooks and Siti Norbaizura Md Zain (2015). Biofilm in the dairy industry. Chapter 8: Biofilm contamination of ultrafiltration and reverse osmosis plants. John Wiley & Sons. Oxford.

2. Siti Norbaizura Md Zain, Steve Flint, Rod Bennett and Tay Hong Soon (2016). Characterisation and biofilm screening of the predominant bacteria isolated from whey protein concentrate. Dairy Science & Technology, 96(3), 285-295. doi:10.1007/s13594-015-0264-z

3. Siti Norbaizura Md Zain, Rod Bennett and Steve Flint (2017). The Potential Source of B. licheniformis Contamination During Whey Protein Concentrate 80

Manufacture. Journal of Food Science, 82(3), 751-756. doi: 10.1111/1750- 3841.13633

5

LIST OF PRESENTATIONS

This work has been presented in part in the following presentations:

Oral Presentations:

1. Siti Norbaizura Md Zain. The role of biofilm development on ultrafiltration membranes in the contamination of whey products. IFNHH Food Division Symposium, Massey University, Palmerston North, New Zealand, 15 November 2013.

2. Siti Norbaizura Md Zain. Identification of bacteria isolated from whey powder. IFNHH Food Division Symposium, Massey University, Palmerston North, New Zealand, 14 November 2014.

3. Siti Norbaizura Md Zain, Steve Flint, Rod Bennett and Tay Hong Soon. Identification of bacteria isolated from whey powder. New Zealand Microbiological Society Conference, Wellington, 18 - 21 November 2014.

4. Siti Norbaizura Md Zain, Steve Flint and Nur Hanizah Kamsani. Biofilm formation of single and mixed strains of thermophilic bacteria isolated from whey powder. Malaysian Society for Microbiology Conference, Terengganu, Malaysia, 6 - 8 December 2014.

5. Siti Norbaizura Md Zain, Steve Flint, Rod Bennett and Tay Hong Soon. Biofilm formation by B. licheniformis isolated from whey protein concentrate 80 powder on two different media and surfaces. New Zealand Microbiological Society Conference, Rotorua, 2 - 5 November 2015.

6

6. Siti Norbaizura Md Zain, Steve Flint and Rod Bennett. Biofilm formation from a co-culture of Bacillus licheniformis and Lactobacillus plantarum on stainless steel in a whey environment. New Zealand Microbiological Society Conference, Christchurch, 14 - 17 November 2016.

7. Steve Flint, Siti Norbaizura Md Zain and Rod Bennett. The role of biofilms in the quality of dairy products in whey processing plants. American Dairy Science Association Annual Meeting, Pittsburgh, Pennsylvania, 25 - 28 June 2017.

Poster Presentations:

1. Rod Bennett, Siti Norbaizura Md Zain, Steve Flint and Tay Hong Soon (2015). Understanding microbial contamination during whey protein concentrate manufacture. Seminar on Emerging Dairy Technologies, Munich, Germany, 16 - 18 September 2015.

2. Siti Norbaizura Md Zain, Mandy Kee and Steve Flint. Isolation and identification of Paenibacillus species from New Zealand dairy product. The New Zealand Institute of Food Science and Technology Conference, Rotorua, 4 - 7 July 2016.

7

ACKNOWLEDGEMENTS

I am indebted to my chief supervisor, Professor. Steve Flint who was also my supervisor during my Master’s study at Massey University in 2008. He was a mentor, role model and counsellor for me during my 4 years of PhD study. His encouragement and guidance have become the backbone of my persistence to excel in my study. My second supervisor, Mr. Rod Bennett was willing to sacrifice his time and thought in this once in a life time process. His expertise in cheese making and whey processing technologies enabled me to understand the process and resolve the questions that arose during this study. I would like to thank my third supervisor, Dr. Tay Hong Soon, for contributing his expertise in microbiology in order to improve the content of this study.

Special thanks to Mr. Eddie Smolinski, Microbiologist at Fonterra, Clandeboye for providing me with the samples, being my industrial mentor and sharing his dairy industrial/process knowledge.

I am grateful to the Universiti Teknologi MARA (UiTM) Malaysia and Ministry of Higher Education (MOHE) of Government Malaysia for providing the scholarship to enable to study at Massey University.

I wish to thank Anne Marie Jackson, Julia Good, Kylie Evans, John Edwards and John Sykes from SEAT for providing the materials that I needed and equipment in the Micro suite lab and also for providing the facilities which enabled this study.

Special thanks to Jordan Taylor and Niki Murray from Manawatu Microscopy Imaging Centre who provided their skills and expertise in taking images using Scanning electron and Transmission electrograph microscopy. With their help, one of my TEM photos (Paenibacillus glucanolyticus spore) was selected as the “Photo of the Day” by American Society for Microbiology (ASM) on their Facebook page on 7th March 2015.

Thanks to Shampa De, Human Nutrition Lab for helping me with the Nanodrop 100 Spectrophotometer instrument.

8

I am blessed with the companionship of my biofilms research team mates; Elham Khanipour, Haoran Wang, Michael Dixon, Jessica Nowak, Shuyan Wu, Dong Zhang and Emmanuel Kyere who made this journey fun and not a lonely path for me.

To my office mates, Nikhila Mary Vijay and Marina Marinae, thank you so much for your encouragement and company during my ups and down.

Not to forget my family, my husband Mohamed Ridzuan, my children; Rifqi, Ariq, Mya and Luqman who were always encouraging me to strive through this tough and meaningful journey. Last but not least, to my mum, Norhayati Hamid, my brother Mohd Nor Azrin bin Md Zain and sister, Siti Nurul Ashikin binti Md Zain for encouraging me to complete my PhD journey. This study would not have succeeded without their understanding and support.

9

LIST OF CONTENTS

ABSTRACT 1

LIST OF PUBLICATIONS 5

LIST OF PRESENTATIONS 6

ACKNOWLEDGEMENTS 8

LIST OF CONTENTS 10

LIST OF FIGURES 15

LIST OF TABLES 17

CHAPTER ONE

1.0 INTRODUCTION 19 1.1 BACKGROUND 20 1.2 RESEARCH QUESTIONS 20 1.3 HYPOTHESES 21 1.4 OBJECTIVES OF THE STUDY 21 1.5 SIGNIFICANCE OF RESEARCH 22

CHAPTER TWO

2.0 LITERATURE REVIEW 23

2.1 MICROORGANISMS IN THE DAIRY INDUSTRY 24

2.1.1 Thermophilic, mesophilic and psychrophilic bacteria 24

2.2 BIOFILM 29

2.2.1 Factors affecting biofilm formation 30

2.2.2 Controlling biofilm formation 35

2.3 WHEY 38 2.3.1 Whey manufacturing process 41

10

2.4 CONCLUSIONS 45

CHAPTER THREE

3.0 ISOLATION AND IDENTIFICATION OF MICROORGANISMS FROM WPC80 POWDER 47 3.1 INTRODUCTION 48

3.2 MATERIALS AND METHODS 49

3.2.1 Source of samples 49

3.2.2 Isolation of bacteria 50

3.2.3 Phenotypic characterisation of isolates 50

3.2.4 Identification by PCR 52

3.2.5 Partial 16S rDNA gene sequencing 52

3.3 RESULTS 53

3.3.1 Isolation of bacteria 53

3.3.2 Phenotypic characterisation of isolates 56

3.3.3 Identification by PCR 62

3.3.4 Partial 16S rDNA gene sequencing 62

3.4 DISCUSSION 66

3.5 CONCLUSIONS 69

CHAPTER FOUR

4.0 ATTACHMENT AND BIOFILM FORMATION BY B. licheniformis 70

4.1 INTRODUCTION 71

4.2 MATERIALS AND METHODS 72

4.2.1 Source of strains 72

4.2.2 Attachment and biofilm screening 72

11

4.2.3 Biofilm formation on SS using three different concentrations of reconstituted whey 73

4.2.4 Biofilm formation on SS using three different media 74

4.2.5 Biofilm formation on SS with effects of individual cations 75

4.2.6 Statistical analysis 75

4.3 RESULTS 76

4.3.1 Attachment and biofilm study of B. licheniformis isolates 76

4.4 DISCUSSION 86

4.5 CONCLUSIONS 90

CHAPTER FIVE

5.0 CHARACTERISATION OF B. licheniformis 91

5.1 INTRODUCTION 92

5.2 MATERIALS AND METHODS 93

5.2.1 Growth at 10°C 93

5.2.2 Lactose fermentation 94

5.2.3 Protease and lipase enzyme production 94

5.2.4 Nitrate conversion 95

5.2.5 Haemolysis on Columbia sheep blood agar 95

5.2.6 Lichenysin synthetase gene (Lch AA) 95

5.2.7 Pellicle formation and Congo red binding assay 96

5.2.8 Heat resistance study at 72°C, 75°C and 80°C 97

5.3 RESULTS 98

5.3.1 Lactose fermentation, protease, lipase, nitrate conversion,

Columbia sheep blood agar haemolysis and lichenysin

synthetase gene A 98

12

5.3.2 Pellicle formation and Congo red binding assay 100

5.3.3 Heat resistance study at 72oC, 75oC and 80oC 102

5.4 DISCUSSION 103

5.5 CONCLUSIONS 107

CHAPTER SIX

6.0 BIOFILM AND SPORE FORMATION BY B. licheniformis IN SINGLE AND CO-CULTURE WITH L. plantarum IN 1% RWPC80 WITH LACTOSE AND MINERALS 108 6.1 INTRODUCTION 109

6.2 MATERIALS AND METHODS 110

6.2.1 Source of the isolates 110

6.2.2 B. licheniformis E30C11 and F30C02 biofilm formation at 1,2,4,8, 12 and 24 h in 1% RWPC80 with lactose and minerals at 37°C 110 6.2.3 Co-culture study between B. licheniformis (E30C11 and F30C02) with L. plantarum biofilm growth at 4, 8, 12 and 24 h in 1% RWPC80 with lactose and minerals 111 6.2.4 Spore formation by B. licheniformis within biofilm in single and co-culture population 111 6.3 RESULTS 112

6.3.1 B. licheniformis E30C11 and F30C02 biofilm growth at 1, 2, 4, 8, 12 and 24 h in 1% RWPC80 with lactose and minerals at 37°C 112 6.3.2 Biofilm of co-culture (B. licheniformis and L. plantarum) at 4, 8, 12 and 24 h in 1% RWPC80 with lactose and minerals at 37°C 113 6.3.3 Spore formation within a co-culture population in a biofilm 115

6.4 DISCUSSION 116

6.5 CONCLUSIONS 118

13

CHAPTER SEVEN

7.0 THE POTENTIAL SOURCE OF B. licheniformis CONTAMINATION DURING MOZZARELLA WPC80 MANUFACTURE 119 7.1 INTRODUCTION 120

7.2 MATERIALS AND METHODS 121

7.2.1 Source of strains 121

7.2.2 Liquid whey samples from pre-UF sites 121

7.2.3 Isolation, identification, characterisation and biofilm formation of B. licheniformis isolates from pre-UF liquid whey samples 121 7.3 RESULTS 122

7.3.1 Isolation, identification and characterisation of B. licheniformis from pre-UF liquid whey samples 122 7.3.2 Biofilm formation by pre-UF B. licheniformis isolates on plastic and SS 125 7.4 DISCUSSION 128

7.5 CONCLUSIONS 130

CHAPTER EIGHT

8.0 FINAL DISCUSSION AND RECOMMENDATIONS 131

8.1 DISCUSSION 132

8.2 CONCLUSIONS 138

8.3 HIGHLIGHTS OF THE STUDY 138

8.4 FUTURE RECOMMENDATIONS 139

REFERENCES 140

APPENDICES 152

14

LIST OF FIGURES

Figure Page

2.1 Image of SS on contact with whey for 24 h 31

2.2 Standard whey content before UF process 39

2.3 The use of whey protein in the food industry 40

2.4 Flowchart of WPC80 manufacture 42

3.1 Morphologies of B. licheniformis cells 57

3.2 B. licheniformis colony morphology on MPCA 58

3.3 SEM image of B. licheniformis cell 63

3.4 SEM image of P. glucanolyticus cell 64

3.5 TEM image of P. glucanolyticus spore 65

4.1 B. licheniformis attachment at 30°C 77

4.2 B. licheniformis attachment at 55°C 78

4.3 B. licheniformis attachment at 30°C, 37°C and 55°C 79

4.4 Graph of 3 different concentrations of RWPC80 81

4.5 Biofilm formation of 33 B. licheniformis on SS 82

4.6 B. licheniformis biofilm formation with cations effect (Ca2+ and Mg2+) 84

15

5.1 Pellicles formation of B. licheniformis in TSB broth 100

5.2 The percentage of Congo red bound for EPS production 101

5.3 The heat resistance of B. licheniformis vegetative cells at 72°C 102

5.4 The heat resistance of B. licheniformis vegetative cells at 75°C 102

5.5 The heat resistance of B. licheniformis vegetative cells at 80°C 102

6.1 Biofilm formation of individual B. licheniformis strain of E30C11

and F30C02 112

6.2 Biofilm formation of co-culture B. licheniformis and L. plantarum

on MPCA 113

6.3 Biofilm formation of L. plantarum on MRSA 113

7.1 Phenotypic characterisation of pre-UF B. licheniformis isolates 123

7.2 Biofilm formation of 7 pre-UF B. licheniformis isolates based

on microtitre plate assay 125

7.3 Biofilm formations of 7 pre-UF B. licheniformis isolates on SS

using 3 different media 126

7.4 The diagram of WPC80 processing by one dairy manufacturing

plant in NZ 127

16

LIST OF TABLES

Tables Page

2.1 The effects of substratum 35

2.2 Description of biosurfactant 37

2.3 Different type of filtration process 43

3.1 Six Mozzarella WPC80 powder samples 50

3.2 Biochemical tests on Bacillus species 51

3.3 Microbial load at 30°C 53

3.4 Microbial load at 55°C 54

3.5 Mesophilic spore counts 54

3.6 List of entire bacteria isolated from each 6 of WPC80 samples 55

3.7 Biochemical test results 59

3.8 API 50CHB test results 60

3.9 Summary of isolates from WPC80 powder samples 62

4.1 Mineral contents of the artificial whey permeate 74

4.2 Biofilm categories determined by O.D absorbance at 570 nm 80

4.3 Summary of frequency of B. licheniformis biofilm 80

17

5.1 Results for B. licheniformis characterisation 99

6.1 Summary of two isolated B. licheniformis strains 110

7.1 Liquid whey samples from pre-UF 121

7.2 Summary of B. licheniformis characteristics 124

18

______CHAPTER 1

______

INTRODUCTION

19

1.1 BACKGROUND Contamination of whey products with bacteria is an economic problem for the dairy industry. Langeveld et al, (1995) noted the high concentration bacterial cells in milk was from their detachment from biofilms into the final product resulting in bacterial counts of >106 per mL, increasing the potential for product spoilage and economic loss. The role of biofilms as the source of the contaminating bacteria in whey products is unknown. Biofilms can form on any surface, but as a large amount of surface area (3500 -7500 m2) is included in the UF membranes used in whey processing, this is one key potential site for biofilm development leading to product contamination.

Whey provides nutrients to enable bacterial growth and the temperatures in different sections of dairy manufacture are suitable for the growth of many different types of bacteria. Growth of bacteria in the dairy liquids (milk and whey) during processing may occur anywhere but the length of time these liquids are at temperatures suitable for microbial growth is limited. Bacterial growth on surfaces (biofilm) is believed to be the main source of bacterial contamination of dairy products (Teh et al, 2015). There is ample opportunity for biofilm development in dairy manufacturing plant and it is likely that the contaminants seen in whey powder originate from biofilms shedding bacteria into the product. In addition, whey protein, along with other organic or inorganic molecules can form a conditioning layer on surfaces that may make them more likely to attract bacteria and start biofilm colonisation.

1.2 RESEARCH QUESTIONS

1) What is the microbial profile of WPC80? a. What are the predominant contaminants? b. What are the characteristics of these contaminants that influence their growth and persistence in whey processing plants?

20

2) What are the relationships between the isolated microorganisms and biofilm formation? a. What conditions favour biofilm formation of the main contaminant (s) in WPC80? b. What is the best test method to examine biofilm formation of these isolates that reflects growth on surfaces in whey processing plants? c. Where is the potential source/location of the contaminant during WPC80 manufacture?

1.3 HYPOTHESES Whey powder quality is affected by the manufacturing process which is related to: 1. Biofilm/ biofouling formation by microorganisms in the whey processing plant. 2. The potential source/location of the contamination is at UF membranes.

1.4 OBJECTIVES OF THE STUDY Microbial biofilms are a major problem in in the dairy industry due to the release of bacteria and enzymes from biofilms into the product. Biofilm growth is supported by the nutritional composition of dairy liquids. Prolonged manufacturing runs and insufficient Clean-In-Place (CIP) increases biofilm accumulation. A better understanding of the biofilms formation of bacteria contaminating WPC and the potential source/location of contamination in a whey processing plant will help the dairy manufacturer in modifying dairy processes to control these bacteria and ensure product quality. Key objectives were:

1. To identify the contaminants of WPC80 powder (Chapter 3).

2. To investigate the attachment and biofilm formation of the isolated bacteria (Chapter 4).

3. To characterise the contaminants (Chapter 5).

21

4. To investigate the co-culture and spore production of the contaminants (Chapter 6).

5. To identify the potential sources of contamination in the whey processing plant (Chapter 7).

1.5 SIGNIFICANCE OF RESEARCH This study was designed to explore the fundamental properties of the bacteria isolated from whey and their ability to form biofilm, produce spoilage enzymes in whey processing plants from either single or mixed cultures and the potential sources of microbial contamination. This knowledge was aimed at helping design strategies to prevent microbial contamination in WPC manufacture.

22

______CHAPTER 2

______

LITERATURE REVIEW

23

2.1 MICROORGANISMS IN THE DAIRY INDUSTRY

There are many different types of bacteria in the dairy industry. Raw milk contains a variety of bacteria from the cow and the milk harvesting process. Throughout the handling and processing of milk into manufactured products, conditions that the milk is exposed to result in changes in the microbial population of the milk. These microbial populations can result in contamination of the product with associated spoilage and undesirable high numbers of bacteria in the final product. Pathogenic bacteria and many of the spoilage bacteria are killed early in the manufacturing process by heat treatment.

2.1.1 Thermophilic, mesophilic and psychrophilic bacteria

Many thermophilic and thermotolerant bacteria survive heat treatment such as pasteurisation and can survive manufacturing processes to contaminate the final product. The common bacteria that survive heat treatment and contaminate dairy products are thermophilic streptococci such as S. thermophilus and various bacillus species such as Anoxybacillus flavithermus (A. flavithermus) and Geobacillus species. Thermophilic are used as hygiene indicators in processed product as they are believed to grow in the manufacturing plant as biofilms, so their presence is an indicator of inadequate cleaning. Furthermore, they are able to produce enzymes and acids that may lead to off-flavours in the final product (Flint et al., 2011b; Burgess et al, 2009; Burgess et al, 2010; Rucket et al, 2004).

Streptococcus species are facultative anaerobic, non-spore forming, catalase- negative, homo fermentative bacteria and have complex nutritional requirements (Flint et al, 1999a). They are lactic acid producing bacteria which can survive high temperatures (Delorme, 2008). They can be found in raw milk (Flint et al., 2011a) and have the ability to survive pasteurisation, hence they are termed thermotolerant. Their optimum growth temperature is around 40°C therefore they cannot be termed “thermophilic” as the in the dairy industry, thermophilic bacteria are those that grow at 55°C. S. thermophilus is also used as a starter to manufacture dairy products especially yogurt and some types of cheese. According to Delorme (2008), this bacterium is used in fermentation due to its ability to rapidly convert lactose into lactic acid resulting in a rapid change in pH, an important part of cheese manufacture. Furthermore, it is able to

24

produce extracellular polymeric substances (EPS) that contribute to the texture of fermented milk products (Delorme, 2008). Even though S. thermophilus can be desirable in the manufacture of some dairy products, it is still a concern when it is able to adhere to surfaces as it can contaminate products that are not supposed to contain high levels of this bacterium, resulting in products exceeding customers’ specifications for microbial levels and associated sensory problems.

One thermophilic bacterium commonly isolated from dried dairy product is A. flavithermus. It is a non-pathogenic bacillus and spore former. The spores are heat resistant (surviving pasteurisation) and the vegetative cells are able to grow up to 65oC for 6 h (Burgess et al, 2009). A. flavithermus is a potential contaminant in milk powder and may deteriorate the quality of the product (Palmer et al, 2010).

Geobacillus species are another group of thermophilic, spore forming non- pathogenic bacteria also commonly found in dairy products. The spores of these microorganisms tend to be somewhat more heat resistant at 100°C with less log reduction (Witthuhn et al, 2011), D-values at 121°C (Dogan et al, 2009), than those of A. flavithermus and the maximum temperature for growth tends to be higher than that for A. flavithermus (Burgess et al, 2010; Seale et al, 2008).

Bacillus species as well as the closely related A. flavithermus and Geobacillus species are common contaminants in dried dairy products. This is because they produce spores that enable their survival. Pasvolsky et al, (2014) reported that Bacillus species contribute to hygiene and sanitation problems since they can form biofilm on the surface of dairy equipment as well as milking pipelines.

Thermophilic bacilli isolated from the dairy industry at 55°C, the temperature routinely used in the dairy industry to isolate thermophilic bacteria, can be divided into two groups: obligate thermophiles and facultative thermophiles. The obligate thermophiles grow only at elevated temperatures of approximately 40 - 68°C (Burgess et al, 2010). Facultative thermophiles belong to the Bacillus genus and are able to grow at both mesophilic and thermophilic temperatures depending on strains (Burgess et al, 2010). Thermophilic bacilli are regarded as indicators of poor hygiene in the dairy industry. Furthermore, thermophiles are potential spoilage microorganisms as they are

25

capable of producing enzymes and acids that may lead to off-flavours in the final product (Teh et al, 2012; Teh et al, 2013; Sadiq et al, 2016b). Thermophilic bacilli are likely to form biofilms in specific unit operations in dairy manufacture where temperatures are 40 – 65oC. These include separators, plate heat exchangers used during pasteurisation, cream heaters in anhydrous milk fat plants and UF plant that operating at hot temperatures (50 -55oC) (Flint et al, 1997).

A study by Yuan et al, (2012) on commercial powder milk in China examined the different species of thermophilic bacilli found in their product. They identified 801 isolates from 22 milk powder samples with 80.5% identified as Bacillus species (B. licheniformis, A. flavithermus and G. stearothermophilus). B. licheniformis was the dominant strain representing 27.8% of the isolates. They also reported that four heat treatments were used to study the heat resistant of 5 B. licheniformis isolates. For the vegetative cells of 4 out of 5 B. licheniformis isolates survived pasteurisation 72°C for 15 s; 3 out of 5 B. licheniformis isolates survived heat treatment at 85°C for 1 min and 2 out of 5 B. licheniformis vegetative cells survived 93°C for 3 min. Spores of B. licheniformis survived all of the three mentioned temperatures. However, both vegetative cells and spores were reduced by 8 log CFU/mL at 121°C for 15 s.

Parkar et al, (2001) showed a variety of thermophilic/thermotolerant bacilli strains including Geobacillus, Anoxybacillus, B. licheniformis, B. coagulans and B. pumilus were able to attach to both SS and milk foulant in similar numbers. Gilmore & Rowe, (1990) noted that some strains of B. licheniformis can produce extracellular substances that appear slimy and sticky and that these may affect the quality of pasteurised cream and milk. Agents targeting the EPS components are frequently reported to induce biofilm dispersion. B. licheniformis excretes an extracellular Dnase (NucB) that rapidly disperses the biofilms formed by both Gram-positive and Gram- negative bacteria (Nijland et al, 2010). B. licheniformis is also associated with a variety of clinical syndromes, such as enteric disease, septicaemia, peritonitis, ophthalmitis and food poisoning. It is also responsible for food spoilage such as ropy bread (Fernandéz- No et al, 2011).

26

Studies by Cook & Sanderman, (2000), Dhakal, (2013) and Sadiq et al, (2016a), reported B. licheniformis was the dominant thermophilic spore former in dairy products where the latter concluded that B. licheniformis contamination occurs in milk and dairy products as many different genotypes that vary in terms of heat resistance and biofilm capabilities. Thermotolerant B. licheniformis is a common species isolated from raw milk and contaminating the dairy production chains with spore production up to 5 log CFU/mL (Crielly et al, 1994). It was identified as the predominant bacteria colonising manufacturing UF plant used for standardising milk used in cheese manufacture (Lehmann, 1995). Growth most likely starts in the pre-heating plate heat exchangers of the UF plant.

Raats et al, (2011) found that raw milk communities did not contain B. licheniformis. They did, however detect Bacillus and Corynebacterium species in 2/3 of the samples. The dominant species they detected were Streptococcus and Lactococcus. Lactococcus species can also dominate in whey UF membranes. Chamberland et al, (2017b) suggesting that these species can survive through many dairy manufacturing processes.

Another raw milk contaminant is Bacillus cereus (B. cereus). This bacterium is ubiquitous in nature, and can be isolated from a variety of processed and raw foods, including milk. However, its presence in foods is not a significant health threat unless it is present in high numbers (>106 CFU/g) or able to grow (Logan, 2011). It has been recognised as a causative agent of food poisoning for more than 40 years producing emetic and diarrhoea symptoms (Fernandéz-No et al, 2011). B. cereus is a spore forming bacterium and a good biofilm former.

Bacillus thermoamylovorans is an emerging spore forming bacilli that is believed to be a threat to dairy product quality, with the ability to survive through heat sterilisation (UHT). B. thermoamylovorans has the ability produce lipolytic and β- galactosidase enzymes are linked to food spoilage (Flint et al, 2017).

Another spore forming genera associated with the dairy industry is the psychrophilic Paenibacillus. Since 1993, the genus Paenibacillus has grown from 11 species to over 26 species (Daane et al, 2002). Paenibacillus are Gram positive, rod

27

shaped, spore formers, motile with petrichious flagella, some of which are facultative anaerobes while others are strict aerobes (Vithanage et al, 2014; Ivy et al, 2012). Paenibacillus are ubiquitous in nature and have been isolated from soil, water, food, faeces, plant materials and diseased insect larvae (Yoon et al, 1998). They produce spores that are heat resistant and able to survive heat treatment such as pasteurisation (De Jonghe et al, 2010). Scheldeman & Goossens, (2004) reported Paenibacillus as the pre-dominant microorganisms (95%) in refrigerated milk at 6oC after 10 days of storage. Paenibacillus can be distinguished from Bacillus species based on their ability to proliferate under refrigeration temperatures to high numbers after an extended storage (Ranieri et al, 2009).

Paenibacillus also were reported able to survive pasteurisation such as high temperature short time (HTST) and low temperature long time (LTLT) (Ranieri et al, 2009). In addition, some Paenibacillus strains are capable of producing spores that can survive ultra-high temperature (UHT) of 140°C to 145oC for 2 s to 5 s as well as retort processes of 110oC to 120oC for 12 to 20 minutes (Burgess et al, 2010).

Many of these bacteria in the dairy industry have been found to produce biofilms on dairy manufacturing plant surfaces (Flint, 1998; Burgess et al, 2010; Teh et al, 2012), and therefore are a potential source of dairy product contamination and quality issues.

28

2.2 BIOFILM

A biofilm is a population of microbial cells growing on a surface and enclosed in an amorphous extracellular matrix (Donlan, 2002). In most environments, microorganisms will grow and develop biofilms as their dominant form and exist on a solid liquid interface. The depletion of dissolved oxygen triggers the formation of floating biofilms. This floating cellular community is called a pellicle (a type of biofilm). Biofilm may consist of either single or multi species microorganisms and represents an important role in many ecosystems (Costerton, 1995; Stewart & Costerton, 2001; O’toole at al, 2000). The formation of biofilms of microorganisms is a protective mechanism for the survival of bacteria in harsh and hostile environments, protecting them from environmental stresses and nutrients depletion (Palmer et al, 2007). Under suitable conditions taking account intrinsic and extrinsic factors, all microorganisms can form biofilms. However, it was noted that some bacteria have a greater tendency than others to form a biofilm. Poulsen, (1999) reported that Pseudomonas, Enterobacter, Flavobacterium, Alcaligenes, Staphylococcus and Bacillus are the most common bacteria associated with biofilms.

Biofilm development starts with the adhesion/attachment of the bacteria to a substrate. This process consists of two phases: the reversible and the irreversible. The reversible phase includes various long distance interactions such as electrostatic, hydrophobic interactions and van der Waals forces. Next, the irreversible phase includes various short range forces such as dipole-dipole, hydrophobic, ion-dipole, ion-ion, covalent bonds and hydrogen bonds (Hood & Zottola, 1995; Kumar & Anand, 1998).

29

2.2.1 Factors affecting biofilm formation

Attachment of bacteria to a substrate is the first stage in biofilm formation. Both the substrate and the bacteria are important in this initial attachment phase The physicochemical properties of the substrate also influence bacterial attachment. Microorganisms attach rapidly on hydrophobic (non-polar surfaces) such as Teflon and polystyrene. Meanwhile, hydrophilic surfaces such as glass and metal are suitable for their attachment (Husmark, 1997). The hydrophobicity of the microbial cell surface is another crucial factor in the adhesion process because of hydrophobic interactions between microbial cell surface and substratum likely to increase with the increasing non-polar properties of one or both surfaces (Wiencek et al, 1991).

Attachment on hydrophobic surfaces, such as microtitre plates, is a good screening tool for biofilm formation however not all microtitre plates are suitable as different types polystyrene amphipathic plasticizing agents are added which affect the results (Donlan, 2002). Polystyrene may also carry net negative charge which may influence adhesion as electrostatic forces are also involved in the adhesion process (Husmark, 1993). SS is another example of a hydrophobic surface. From studies by Flint et al, (1999a; 1999b) on the removal of thermo-resistant streptococci, S. thermophilus was able to adhere to SS and withstand various cleaners and sanitisers especially in warm zones of dairy manufacturing plants, where Streptococci attach directly to SS. The dairy industry has used SS for more than 60 years in almost all milk handling systems (Marchand et al, 2012). SS is the most common material used in food processing plant surfaces as it is corrosion resistant and easy to clean as described by Holah & Gibson (2000). The properties of SS such as grade (304 or 316) and surface finish do impact the SS hydrophobicity with 316 SS more hydrophilic than 316 SS with 2B finish (Teh et al, 2015). Further modification of SS surfaces might be useful to reduce or mitigate bacterial attachment and biofilm formation. The surface properties of SS show some crevasses that bacteria can be trapped in (Figure 2.1). This microscopic appearance is playing a vital role for bacterial attachment or cleaning or both.

30

Figure 2.1 Image of SS on contact with whey for 24 h which. There are deposits on the SS surface (presumably protein) that may attract bacterial attachment. Note the crevices and pits on the surface which can trap microbial cells.

Bacterial attachment occurs best on surfaces that are hydrophobic, rough and coated with conditioning films. A conditioning film is a layer of organic material that can alter the conditions on the substrate, making it more or less likely to support bacterial attachment (Tang et al, 2015; Teh et al, 2015). Zaky et al, (2012) studied the role of conditioning films in the early stage of membrane fouling. In addition, nutrient concentration, temperature and flow velocity do enhance the attachment process. (Anand et al, 2014; Yang et al, 2012). Hydrophobic interaction depends on the hydrophobic properties of the cell surfaces and hydrophobic properties of the substrate. Heat treatment (80°C for 10 min) induces Bacillus spores to be more hydrophobic (Husmark, 1993). Hydrophobic interaction increased with temperature. The more

31

hydrophobic the substrate and the cell surface the higher the percentage of adhering spores. The contaminating microorganisms’ adherent spores or bacteria develop into biofilms on every surface including SS, floors, belts or rubber seals in the food industry (Costerton, 1995; Kumar & Anand, 1998).

Both the vegetative cells and the spores of Bacillus species are involved in the formation of a biofilm. The spore surfaces are hydrophobic due to proteins on their outer surface (Wiencek et al, 1990). The high relative hydrophobicity of Bacillus spores is reflected in a high degree of attachment and the hydrophobic interactions increase proportionally with temperature. Spores from different Bacillus species vary in their attachment to solid surfaces. Husmark, (1993) discovered that attachment was higher in the spore state than the vegetative state in five different Bacillus species and B. cereus is the most hydrophobic species. The attachment process is quite rapid and within 1 hour, a maximum level of adhesion is reached for specific spore concentration. In general, the mechanism of microbial attachment and growth as biofilm on any surface is a very complex process with internal and external factors affecting the process. Although spores, if present, dominate in the attachment process, Faille et al, (2014) found no correlation between biofilm and sporulation within biofilms of Bacillus species.

Seale et al, (2008) noted, “the outer layer of a Bacillus subtilis spore consists of protein and carbohydrates while in species such as Bacillus cereus and Bacillus anthracis, the spores exhibit an additional layer, known as an exoporium comprising proteins, carbohydrates and lipids”. The factors influencing spore attachment to surfaces are primarily hydrophobicity, electrostatic interactions and surface polymer conformation (Seale et al, 2010). Surface polymer conformation controls the attachment of thermophilic spores such as Geobacillus and A. flavithermus. Surface hydrophobicity enhances bacterial attachment with spores being more hydrophobic and hence more likely to attach compared with vegetative cells (Wiencek et al, 1991; Husmark, 1993).

Another factor that contributes to attachment of bacteria to surfaces is cell motility. Korber et al, (1989) reported motile strains of Pseudomonas fluorescens (P. fluorescens) attach rapidly and in higher numbers compared to non-motile strains in

32

continuous flowing conditions. Non-motile P. fluorescens strains appeared to be slow biofilm formers as they are unable to re-colonise and seed the empty areas of the substratum compared to motile strains. They concluded that bacterial appendages such as flagella play an important role in attachment and biofilm formation by overcoming the repulsive electrostatic forces associated with the substratum. Lemon et al, (2007) studied the Listeria monocytogenes (L. monocytogenes) with flagella biofilm formation in which flagella helped the L. monocytogenes motility by propelling them towards the surface to aid attachment in static conditions. Meanwhile in continuous laminar flow conditions, L. monocytogenes without flagella resulted lower attachment in the early stages of biofilm development however the biofilm formation increased with time. Abee et al, (2011) reported the same observation on biofilm formation by B. cereus and the role of flagella to aid the motility of B. cereus in the attachment process.

Biofilm will grow with time and there will be an increase in microorganisms encapsulated in a polymer matrix. Diffusion through this matrix becomes a major factor when determining the structure of the biofilm. The production of EPS by bacteria during the biofilm formation plays many functions. The functions including facilitation of the initial attachment of bacteria to a surface, maintaining the formation of a micro colony and biofilm structure and enhanced resistance of the bacteria in the biofilm to environmental stress and anti-microbial agents (Poulsen, 1999; Yang et al, 2012). EPS are responsible for filtration membrane fouling by irreversibly binding to the membranes. Tang, (2011) quoted that “EPS enhance the survival and robustness of the biofilm microorganisms by forming a chemically reactive diffusion transport barrier for the bacterial cells, impeding convective flow and slowing the penetration of biocide into the biofilm”.

EPS acts as protective layers in holding cells from hostile environments. Although there is evidence of biofilm formation from single species, most biofilms in nature are consist of multi species of microorganisms (Flemming & Wingender, 2010; Mielich-Süss & Lopez, 2015). Van Dyk et al, (2012) reported that different carbon sources (glucose, arabinose, sucrose, and xylose) and nitrogen sources did affect the production of B. licheniformis EPS production.

33

Quorum sensing (QS) is also believed to be involved in biofilm development. QS is cell-to-cell signalling which provides communication between bacterial cells to form mature biofilm (Anand et al, 2014). The QS-coordinated process is achieved by producing, releasing and detecting small signal molecules known as autoinducers (AI). The AI increased with the increasing bacterial cell density. The regulator proteins are triggered once the concentration of AI is at its peak, leading to transcription of QS- regulated genes resulting in changes in bacterial properties. QS has a role in biofilm formation and dispersion. Inhibition of QS systems of microorganisms appears to be a promising method for controlling microbial attachment and membrane fouling (Xiong & Liu, 2010)

A summary of the factors affecting biofilm formation on any surface is presented in the Table 2.1 which includes properties of the substratum, bulk fluid and cells.

Table 2.1 The effects of the substratum, conditioning films forming on the substrate, hydrodynamics of the aqueous medium, characteristics of the medium and various properties of the cell surface on biofilm formation. (Adapted from Donlan, (2002) and Simões et al, (2010)).

34

Properties of the Properties of the bulk Properties of the cell substratum fluid Texture or roughness Flow velocity Cell surface hydrophobicity

Hydrophobicity pH Flagella / Fimbriae

Conditioning film Temperature Quorum sensing

Cations Extracellular polymeric substances

Presence of antimicrobial Spores agents

There are 5 key factors that affecting biofilm formation in the food industry (Teh et al, 2015); 1) Types of microorganisms. 2) Types of product being processed. 3) The operation conditions of the manufacturing plant (length of operation and temperature). 4) Types of surfaces. 5) Cleaning and sanitising regime.

2.2.2 Controlling biofilm formation

Increasing numbers of bacteria in food from a manufacturing plant can be due to biofilm growth and release of bacteria from the biofilm during processing and is exacerbated by an accumulation of residues from incomplete cleaning between production runs (Flint, 1997; Burgess et al, 2010; Murphy et al, 1999; Parkar et al, 2003). Control of microbial colonisation within processing lines in the dairy industry requires CIP procedures. Sodium hydroxide (NaOH) at 1-2% w/v at temperatures > 70°C is widely used in the food and beverage industries (Faille et al, 2013). This procedure has proven effective in dissolving both proteinaceous and fatty soils as well as having emulsifying properties.

35

Anand et al, (2014) proposed the use of surfactants together with enzymes and chelating agents to prevent bacterial attachment and improving cleaning process. Tang et al, (2010) recommended QuatroZyme which is composed of mixed enzymes to be used as a more effective cleaning system than the standard caustic and acid cleaning systems generally used for membrane processing plants.

Commercially produced microbiological products such as bacteriocin (nisin) and biosurfactant (subtilin) can be used to control biofilm. The biosurfactant has anti- adhesive properties that can prevent the attachment of cells and spores to surfaces. Busscher et al, (1996) used biosurfactants produced by thermophilic dairy streptococci (S. thermophilus) for fouling control on heat exchanger plates in pasteurisers. Earlier, Busscher et al, (1990) reported a biosurfactant layer on the surface to which freshly cultured cells did not adhere. This layer was due to biosurfactants produced by the adhering cells themselves. Such a biological anti-adhesive coating might be a potential technique for the control of microbial fouling in the dairy industry. Another role of biosurfactants in a biofilm is their contribution to surface activity (surface tension) in the, early stages of micro-colony formation, aiding surface associated bacterial migration preventing colonisation of channels within biofilms and taking a role in biofilm dispersion (Flemming & Wingender, 2010; Pamp & Tolker-Nielsien, 2007).

A biosurfactant that is produced by B. licheniformis, is lichenysin. Some information on lichenysin is presented in Table 2.2. Lichenysin synthesis is coded by the lch A operon (Ron & Rosenberg, 2001). The production of lichenysin may be regulated by cell density, which sparks several quorum sensing control elements either stimulate or inhibit its production. Lichenysin A is produced by B. licheniformis (Yakimov et al, 1995; 1997) and Lichenysin B is produced by B. licheniformis JF-2 (Lin et al, 1994).

36

Table 2.2 Description on biosurfactant produced by B. licheniformis identified as lichenysin.

Properties Description Source Heat x Lichenysin A is a Mikkola et stable lipopeptide lactone that al, 2000 can be produced aerobically and anaerobically (100°C, 20 mins). Surface x Lipopeptide high surface Rivardo et tension activity and can reduce al, 2011 surface (biofilm) colonisation by E. coli. Anuradha, x Lower surface tension of 2010 water for 72 to 27 mN/m. Emulsion/ x Secondary metabolites Anuradha, surfactant that solubilise 2010 hydrophobic substrates, and regulation of attachment-detachment of microorganisms from surfaces. x Act as antimicrobial and Coronel- pseudo-surfactant León et al, (synergistic effect). 2017

Bacillus species producing biosurfactant in biofilms that reduce the surface tension between the cell surface and the substrate reduce biofilm formation. Faille et al, (2014) noted that Bacillus biofilms adherent poorly to SS in their lab and were even susceptible to detachment during a single rinsing procedure, believed to be due to

37

surfactants produced by these bacteria. A similar observation was made by Wijman et al, (2007).

The formation of biofilm in different parts of dairy manufacturing plant is likely to be influenced by the types of contaminants present that may enhance or reduce the potential for biofilm formation. Some parts of dairy manufacturing plant are likely to be influenced by biofilm formation more than others. As whey originates from the result of several manufacturing processes, this is likely to be influenced by microbial contamination more than other dairy products.

2.3 WHEY

Whey was traditionally a waste product from the manufacture of cheese or casein that needed to be discarded. Whey is difficult to dispose of or use due to its unfavourable lactose-to-protein ratio and high biological oxygen demand of 30 000 – 50 000 ppm (Madaeni & Mansourpanah, 2004). The need to reduce the impact on the environment plus the realisation that whey contains components of value has resulted in the manufacture of numerous whey products. Whey is the liquid remaining after the production of cheese or the removal of fat and casein (80% of the proteins) from milk (Macgibbon, 2014). The worldwide trade of whey ingredients (all types combined) were estimated at 1.5 million metric tonnes in 2013 with European Union the leading exporters (Lagrange et al, 2015). Most of the whey (92%) is cheese whey, the liquid remaining from the production of cheese. Whey contains about 50% of the nutrients present in milk, comprising milk sugar (lactose), serum proteins (whey proteins), minerals, a small amount of fat, and most of the water soluble minor nutrients from milk such as vitamins (Zadow, 1992; Smithers, 2008) (Figure 2.2).

38

Minerals, 0.5, 1%

Fat, 0.3, 0%

Lactose , 4.9, 5% Protein, 0.8, 1% Water Lactose Fat Protein Minerals

Water, 93.5, 93%

Figure 2.2 The standard whey content before the UF process. Information from Wisconsin Centre for Dairy Research, (http:www.cdr.wisc.edu)

The whey constitutes of 90% of the original milk volume with 50% of the original milk components remaining when the casein is removed (Onwulata & Huth, 2008). Whey protein is concentrated and spray dried to provide a product that can be used as an ingredient in many foods (Figure 2.3). The most abundant component of whey after water is the milk sugar lactose can be considered as having nutraceutical roles besides being the main source of energy for the newborn (Morr & Ha, 1991; Morr & Ha, 1993). Lactose is now in very strong demand for protein standardisation in milk powder manufacture. Its disposal is no longer an issue in New Zealand.

39

Confectionery

Dietetic Meat/ fish foods product

Whey protein in food Infant products Bakery formula products

Nutraceuticals Dairy products

Figure 2.3 The use of whey protein in the food industry

WPC prepared by the UF process of and drying of whey contains 68.4% β- lactoglobulin, 21.3% α-lactalbumin and 10.3% serum proteins (Delaney, 1976). The lactose, salts and much of the water are removed during UF and dialysis. The remaining water is removed by evaporation and spray drying. The ratio of individual proteins in any ultrafiltered WPC will be influenced by factors such as:

1) type and source of whey.

2) rejection characteristics of the UF membrane used.

3) degree of concentration achieved.

Whey became of interest to food processing industries as its physical and chemical properties, especially its protein, have unique nutritional qualities. In addition,

40

the undenatured whey proteins have unique functional properties. The major whey product is WPC (Zadow, 1992). WPC can be defined as the product derived from milk whey containing 50, 65 or 80% native proteins (Luck et al, 2013). WPC contains higher tryptophan (non-polar) amino acid and cysteine compared to skim milk but falls below Food and Agriculture Organization (FAO) whole egg reference protein values for valine, tyrosine, phenylalanine and methionine content (Delaney, 1976). The high percentage of lysine, tryptophan, methionine and cysteine in WPC should be noted. However, phenylalanine and tyrosine are limited in gel filtered and ultra-filtered WPC. Low mineral content is desired in WPC because of the demand of certain food formulations such as infant formula and healthy food products. The vitamin content of

WPC is rich in B12 and folic acid. The other water soluble vitamins of whey are in the free form is released through the UF membrane during WPC manufacture.

2.3.1 Whey manufacturing process

A schematic diagram of the whey protein concentrate manufacturing process at one dairy plant in New Zealand is presented in Figure 2.4. The first step is clarification to remove particles of cheese or any precipitated proteins and possibly some bacterial contamination. Thermalisation reduces the microbial load and chilling slows any further microbial growth during UF and dialysis. The thermaliser heat plate exchanger used in this manufacturing plant has temperatures from 36oC to 77°C (Appendix 1). The separator used to remove fat from whey. Once concentrated to approximately 50% of its original volume, and the lactose and salts have been removed by dialysis, the remaining protein solution is concentrated by evaporation by another 50% before spray drying.

41

Mozzarella Raw Clarifier Thermaliser Separator Chiller Ultrafiltration/ process whey Diafiltration balance tank 40°C 14°C 35°C 36-77°C 40-44°C 8°C 10°C

Dryer Concentrate Evaporator Evaporator (air inlet) handling preheat

>200°C 40°C 51-53°C 25°C

Figure 2.4 A flowchart of the whey manufacture processes.

42

A key part of the manufacturing process is UF. This separation takes place through semipermeable membranes, using a hydrostatic pressure gradient as the driving force. UF is molecular sieving on a commercial scale in the manufacture of whey products. UF is a pressure-driven flow-dependent filtration process in which porous membranes are used to separate the components of a solid-liquid mixture on the basis of shape and size and in some instances, charge. When the pressure gradient is applied across the membrane, the liquid is forced to flow through the pores to the low-pressure side, transporting any components that are smaller than the size of the membrane pores. The pressure gradient required for UF is generally low and typically around 500 Kpa. The pore dimensions are typically in the range of 1 to 100 nm diameter. The molecular weight cut-off in the range from 10 000 – 50 000 Daltons used for the fractionation of whey protein in UF membranes. Low molecular weight solids such as lactose, minerals and water will pass through the membrane as permeate, while proteins and residual fat are rejected and stay inside as retentate (Mohammadi et al, 2002; Zadow, 1992) (Table 2.3).

Table 2.3: Different types of filtration process and their function (Zadow, 1992). Type Pore size Components retained Molecular weight of (nm) component (kDa) MF 200-4000 Bacteria, casein micelles, fat globules 100-500 UF 20-200 Whey proteins 1-100 NF < 2 Lactose 0.1-1 RO < 2 Ions < 0.1

Whey separation by UF is normally performed at temperatures below 55°C with an inlet pressure around 300 kPa and a membrane pore size of 250 nm. Whey retentate consists of protein, fat and insoluble salts, lactose, soluble mineral and much water (Onwulata & Huth, 2012). Initially UF plants operated at 50°C for reasons of improved flux and microbiological control. Since the advent of spiral wound membranes with a high surface area to volume ratio, it has become economic to operate plants at 10 °C for improved microbiological control and increased membrane life.

43

Processing operations (pasteurisation, pre-heating, evaporation and spray drying) may result in Maillard type browning and concomitant reduction in available lysine content. The UF step has little adverse effect on the available lysine content as UF is a gentle, low temperature process involving no phase change.

Higher degrees of whey fractionation result in the viscosity of the retentate becoming too high to pump. A process called “Diafiltration” involves adding water during UF to remove salts and lactose thus solved this problem. The composition of a WPC depends on the properties of the membrane, the duration of the filtration process and the use of water (Diafiltration). The UF process is able to produce all types of WPC, with protein contents ranging from 25% to 80% total solids (De Wit, 2001).

Milk fractionation using different pore sized membranes can separate the casein micelles and bovine serum albumin leaving whey serum as the retentate. Further separation of the whey occurs using UF which separates the whey protein consisting of α-lactalbumin and β-lactoglobulin as the retentate and leaves the lactose and peptides as permeate. Nanofiltration is used to fractionate and concentrate the whey protein into its individual components (of α-lactalbumin and β-lactoglobulin). Meanwhile reverse osmosis is used to concentrate the lactose for drying (Morr & Ha, 1993).

Membrane fouling limits the operation of UF membrane plants. Fouling is defined as the attachment and growth of microorganisms and irreversible collection of materials trapped on the membrane surface which results in a flux decline. Fouling of UF membranes in the dairy industry is mostly due to the colonisation of microorganisms, and residual protein, fats and minerals on the membrane surfaces either at its pore openings or within its pores (Mohammadi et al, 2002; Mohammad et al, 2012; Chamberland et al, 2017a). Zaky et al, (2012) defined abiotic and biotic fouling on membranes as:

1) Abiotic fouling: conditioning layer: fouling deposition of organic, salts, dead and cell debris from the feed water on the membrane surface.

2) Biotic biofouling: fouling due to microbial growth from active cells on the membrane surface.

44

Whey is a foulant for membranes in dairy plants with the foulant being mainly made up of protein. In whey, β-lactoglobulin is the most well-known foulant (Fouladitajar et al, 2014). Inorganic salts, including calcium phosphate also contribute to fouling. This salt acts as a binding bridge between the membrane and the protein, leading to hydraulic resistance of the protein layer (Madaeni & Mansourpanah, 2004; Pattaranawik & Leiknes, 2011; Herzberg & Elimelech, 2007; Laborie et al, 1998; Liu et al, 2012). However, biotic fouling such as microbial growth on the membrane surface may result in poor quality product. Studies on biofilm formation on UF membranes by Tang, (2011) showed Gram negative Klebsiella oxytoca were dominant and believed to have originated from water used in the cleaning of the plant and dialysis. Chamberland et al, (2017b) studied the biofilm formation on UF membrane operated at 10°C and found that Lactococcus spp were the dominant contaminant.

2.4 CONCLUSIONS

Microbial contamination of dairy products is controlled by heat treatment such as pasteurisation. However, some thermophilic and mesophilic organisms as well spore forming bacteria can survive and contribute to the spoilage of the milk and dairy products. These bacteria include S. thermophilus and spore-forming bacteria Anoxybacillus, Geobacillus and Bacillus species (Flint, 1998).

Dairy product contamination is enhanced by the growth of biofilms in the manufacturing plant. Biofilms develop from bacteria that survive heat treatment and contaminate any surface where the conditions are suitable for bacterial growth. The contribution of biofilms of these thermo-resistant bacteria to the contamination of whey products is unknown.

Biofilms reduce heat transfer in heat exchangers and cooling towers. The EPS matrix provides protection and resistance to antimicrobial treatments and contributes to a higher transformation frequency (transfer of DNA) among the bacteria comprising the biofilm than planktonic cells (Vlamakis et al, 2013; Flemming & Wingender, 2010). The contribution of different stages of WPC manufacture to contamination of WPC through biofilm development is unknown.

45

Fouling, including biofilm, results in an increase in operational costs, due to an increased energy demand, additional labour for maintenance, cleaning solution costs, shorter membrane life and limitations in the amount of product that can be processed. Heat exchangers or/and thermaliser are cleaned every day by CIP as fouling with whey protein occurs rapidly and reduces heat transfer (Bansal & Chen, 2006). Even this, however, is not always effective in controlling microbial contamination of WPC.

46

______CHAPTER 3

______

ISOLATION AND IDENTIFICATION OF MICROORGANISMS FROM WPC80 POWDER

47

3.1 INTRODUCTION

Whey is the liquid remaining after the production of cheese or the removal of fat and casein (80% of the proteins) from milk, and contains predominantly Gram- positive organisms from the starter population used in cheese and casein manufacture (lactic acid bacteria) or thermo-resistant species such as spore-forming Bacillus species (Schreiber, 2001). There are some starter bacteria, such as S. thermophilus and thermo-resistant Bacillus species can tolerate heat treatment such as thermalisation, pasteurisation and evaporation used in the manufacture of WPC. Whey contains about 50% of the nutrients present in milk, comprising milk sugar (lactose), serum proteins (whey proteins), minerals, a small amount of fat and most of the water soluble minor nutrients from milk such as vitamins (Smithers, 2008; Zadow, 1992). This whey is selectively concentrated by UF and evaporation then spray dried to provide a high protein product that can be used as an ingredient in many foods.

Whey, formerly a waste product, became of interest to the food processing industries due to its physical and chemical properties, especially its protein and nutritional qualities. The major whey product is WPC (Zadow, 1992). It can be defined as the product derived from milk whey containing 50%, 65% or 80% native proteins (Luck et al, 2013) and called WPC50, WPC65 and WPC80, respectively. The individual protein composition in ultra-filtered WPC is about 68% β-lactoglobulin, 21% α-lactalbumin and 10% serum proteins. This is approximately the same as present in whole milk (Delaney, 1976). The amino acid composition of WPC and skim milk were compared and the results showed WPC contains higher tryptophan and cysteine. This is due to the activity of starter bacteria in the manufacture of cheese whey (Macgibbon, 2014).

The microflora in WPC varies depending on the type of whey with sweet whey showing a higher prevalence of thermophilic and mesophilic spore formers compared with acid whey (Watterson et al, 2014). Aerobic spore formers of particular concern in dairy products include the psychrotolerant Paenibacillus species, mesophilic Bacillus species; for example, B. licheniformis, B. subtilis and B. pumilus and thermophilic A. flavithermus and Geobacillus species in a group that is not well defined (Burgess et al, 2009; Burgess et al, 2010; Watterson et al, 2014). Spore

48

formers found in the dairy industry are the resistant forms of Bacillus and Clostridium species that withstand heat and chemical treatment that is used to control most bacteria in dairy manufacture.

Generally, the microbial contaminants in products such as WPC are recorded as either thermophilic or mesophilic bacteria in routine testing. Dairy manufacturers have specifications for these groups of bacteria but there is no requirement for further testing to identify the bacteria comprising these groups. The isolation and identification of these bacteria could help in determining the effect on product quality and the risk of the manufacture of unacceptable product and the risk of product spoilage. The study in this chapter is aimed to isolate and identify the contaminants in the six Mozzarella WPC80 powder samples.

3.2 MATERIALS AND METHODS

3.2.1 Source of samples

Six Mozzarella WPC80 samples from different manufacturing runs were obtained from one dairy manufacturing site in New Zealand. The samples in powdered form were received in 100 g foil lined paper pouches used to store samples from the manufacturing process for testing. The Mozzarella WPC80 samples are listed in Table 3.1. Six different Mozzarella WPC80 samples provided due to reported high bacterial counts were analysed to obtain the total microbial load using the aerobic plate count technique as used in the dairy industry to monitor the quality of the product.

49

Table 3.1 Six Mozzarella WPC80 samples from one dairy company in New Zealand. Samples Details A Cypher: CX15 B Cypher: CX06 C Cypher: CX07 D Cypher: CX09 E Cypher: CX17 F Cypher: CX22

3.2.2 Isolation of bacteria

Ten grams of each WPC80 samples were diluted in 90 g of sterile 0.1% peptone water (pH 7.0) and homogenised using a peristaltic blender for 1 min prior to serial 10-fold dilution to 1 x 10-6 in 0.1% peptone. The pour plate technique was performed using milk plate count agar (MPCA) (Merck, BDH, Palmerston North, New Zealand) in duplicate to obtain numbers in the range of 30 - 300 colonies per plate. The agar plates were incubated aerobically at 30°C for 24 h for mesophilic bacteria and at 55°C for 18 h. Two to three single isolated colonies of different colony morphology were taken from the plates of each Mozzarella WPC80 samples and restreaked onto MPCA to obtain pure cultures. Pure strains were streaked on beads in glycerol and stored at -80°C (Microbank, Pro-Lab Canada). Mesophilic spores from the samples were determined by adding 1g of each WPC80 samples into 99 mL sterile distilled water and heated at 85°C for 10 mins. Serial dilutions were performed until 10-5 and plated on MPCA in duplicate. The plates were incubated for 48 h at 30°C.

3.2.3 Phenotypic characterisation of isolates

The pure cultures grown on MPCA were screened using microscopic observation of Gram stained cells. The motility tests were done by the hanging drop technique, motility agar and observed under light microscopy (Olympus, USA). Some of the isolates samples were sent to the Manawatu Microscopy Imaging Centre for scanning electron microscope (SEM) and transmission electron microscope (TEM) images. These were used as additional confirmatory tests for some of the isolates,

50

selected on the basis of colony morphology as presumptive B. licheniformis. Supportive information on the identification of the bacterial isolates was obtained by biochemical testing using the conventional culture methods in selective growth media or miniaturised API 50CHB system. The Bacillus species were isolated on MYP agar (mannitol-egg yolk polymyxin; Oxoid) for presumptive B. cereus identification. After incubation of plates for 24 to 48 h at 30°C, characteristic pink colonies, non- fermenting mannitol and surrounded by a zone of white precipitation of lecithin were recorded (Appendix 2). Further identification was done by biochemical tests using the following media in the Table 3.2 below, incubating at 30°C for 24 h. A 24 h culture was prepared and 1 mL of the culture was pipetted into the API 50CHB inoculation medium and subsequently pipetted into each tube but not the cupule. The strip was incubated at 30°C for 48 hours and observed after 24 and 48 h. After 48 h of incubation, the results were coded and interpreted using the API website (bioMérieux).

Table 3.2 Biochemical tests done to help identify unknown Bacillus species.

No Media Manufacture Purpose

1 Mannitol egg-yolk Becton Dickson, USA B. cereus identification

polymyxin agar

2 Simmons citrate OXOID, England Citrate utilisation

3 Gelatine agar OXOID, England Gelatine production

4 Nutrient broth with Becton Dickson, USA NaCl tolerance

7% NaCl

51

3.2.4 Identification by PCR

Gram positive Bacilli with colonies indicative of B. licheniformis, were further tested using species specific PCR targeting the gyrase B gene (Huang et al, 2012) using one set of species specific primers Blich-F1 5’- AKACGGAAGTGACGGGAAC-3’ and Blich-R1 5’- AGAAACTTTTCRAGCGCTT-3’. The DNA was extracted by boiling the bacterial culture. PCR amplification involved 20 μL Master mix (5 Prime MasterMix-100 Rxns GmbH, Germany) consisting of dNTPs, magnesium chloride and Taq DNA polymerase, 24 μL UltraPure™ DNase/RNase-Free Distilled water (Invitrogen), 1 μL forward primer, 1 μL reverse primer and 4 μL DNA template (unknown culture) to achieve a final volume of 50 μL. Amplification conditions were as follows; denaturing step at 94°C for 7 min followed by 35 cycles with denaturation (49°C for 60 s), annealing (55°C for 60 s) and extension (72°C for 60 s) followed by a final extension at 72°C for 15 min. All PCR assays were carried out on the Techno thermal cycler (TC-400, Total Lab Systems, Auckland, New Zealand). The PCR products were visualised (E-Gel iBASE™, Invitrogen) using pre-made 2% agarose electrophoresis gel (E-Gel ® EX with SYBR Gold II) then visualised under UV Transilluminator, UVP, Inc. (Chromato-Vue, San Gabriel, California, USA) and UVITEC (Cambridge, UK).

3.2.5 Partial 16S rDNA gene sequencing

For the colonies that were not indicative of B. licheniformis and for those isolates that did not produce a positive result with the species-specific primers for the gyrB gene of B. licheniformis, universal primers Bac27F 5’- AGAGTTTGATCCTGGCTCAG-3’ and U1492R 5’- TACGGCTACCTTGTTACGACTT-3’ were used to amplify a 1000bp part of the 16S rDNA genome for sequencing, using the conditions in Flint et al (1999b). Prior to sequencing, the PCR products were purified using a Zymo DNA Clean & Concentrator TM-5, USA kit and sent to UV/VIS spectrophotometry to check the DNA concentration (1 μL – 2 μL) using the NanoDrop 1000 Spectrophotometer (Thermo Scientific, USA). Sequencing was done at the Massey University Sequencing

52

Unit using the BigDye terminator v3.1 cycle sequencing kit and the results analysed using Seqmatch and BLAST GeneBank.

3. 3 RESULTS

3.3.1 Isolation of bacteria

Six different Mozzarella WPC80 samples were analysed to obtain the total microbial load using the total plate count technique. Two temperature profiles (30°C and 55°C) were selected for incubation to isolate both mesophilic and thermophilic microorganisms. The predominant bacteria in the WPC80 samples were mesophilic bacteria (Table 3.3) with total mesophilic bacterial counts close to or exceeding the specification limits. The Australia New Zealand Food Safety Authority guideline for the mesophilic plate count for dried milk including whey powder is 5 x 104 CFU/g (acceptable level) and 2 x 105 CFU/g (maximum). The thermophilic bacterial counts were all < 103 CFU/g (Table 3.4). There is no guideline for thermophilic plate count. Mesophilic spore counts mostly below 100 CFU/g except in Sample A which did not yield any positive result. The results are given in Table 3.5

Table 3.3 Microbial load WPC80 samples incubated at 30°C for 24 h determined by plate counting on MPCA.

Sample Total plate count (CFU/g) A 5.5 x 104 B 5.9 x 105 C 1.0 x 104 D 7.3 x 104 E 1.2 x 105 F 3.3 x 104

53

Table 3.4 Microbial load from WPC80 samples incubated at 55°C for 18 h determined by plate counting on MPCA. Sample Total plate count (CFU/g)a A < 102 B < 103 C < 103 D < 103 E < 102 F < 103 a Spreading colonies for the thermophilic aerobic plate count meant no accurate counts could be obtained for this test.

Table 3.5 Mesophilic spore counts determined by plate counting on MPCA from WPC80 samples incubated at 30°C for 48 h. Sample Spore counts (CFU/g)b A 0 B < 100 C < 100 D < 100 E < 100 F < 100 b Spreading colonies after 48 h of incubation.

The list of isolated bacteria from six WPC80 powder is listed in Table 3.6. Mozzarella WPC80 samples were labelled A, B, C, D, E, and F and bacteria isolated from each sample was labelled according to their cypher sources.

54

Table 3.6 List of the bacteria isolated from Mozzarella WPC80 samples and their basic physical properties.

WPC80 Number of Gram staining Incubation Sample Isolates temperature G +ve G -ve bacilli 30°C 55°C bacilli

A 30C01 + + 30C11 + + 30C21 + +

B 30C11 + + 30C12 + + 30C13 + + 30C14 + + 30C21 + + 30C22 + + 30C24 + + 55C01B + + 55C02 + + 55C11 + +

C 30C01 + + 30C11A + + 30C11B + + 30C21 + + 55C01 + + 55C02 + + 55C11 + + 55C12 + +

D 30C01A + + 30C01B + + 30C02 + + 30C03 + + 30C04 + + 30C12 + + 30C14 + + 55C01 + + 55C02 + + 55C03 + + 55C04 + + 55C11 + +

E 30C01 + +

55

30C03 + + 30C04 + + 30C11 + + 30C12 + + 30C21 + + 30C22 + + 55C01 + + 55C02 + +

F 30C01 + + 30C02 + + 30C11 + + 30C14 + + 55C01 + + 55C02 + + 55C11 + + 55C12 + +

3.3.2 Phenotypic characterisation of isolates

The isolates obtained from the six WPC80 samples were predominantly Bacillus species with rod shaped purple-violet colour on Gram staining, as expected from colony morphology. All of the isolated bacteria were Gram positive, mostly spore formers and motile based on light microscopic observation. (Olympus, USA) (Figure 3.1).

The colonies formations of B. licheniformis on MPCA were mucoid, large opaque, adherent, wrinkled and irregular edges (Figure 3.2). The isolates were retrieved from MPCA and incubated at the appropirate temperature (30°C or 55°C). The growth indicated that isolates were facultative anaerobes (the colonies grew best within the solid agar) and others were strict aerobes (the colonies grew best on the agar surface). Although some of the isolates were observed to be facultative anaerobes, their growth in strict anaerobic conditions was not tested.

The summary of biochemical tests is listed in Table 3.6. Biochemical testing was performed to identify the isolates prior to bio-molecular testing. Supportive information on the identification of the bacterial isolates was obtained by conventional culture methods in selective growth media for Bacillus (Table 3.7) followed by biochemical testing using the miniaturised API 50CHB (bioMérieux) (Table 3.8).

56

A

B

Figure 3.1 Image of E30C11 strain (A) and D30C02 strain (B) shows the cell the morphology typical of Bacillus species isolated from WPC80. Some were spotted within the cell (B).

57

A

B

Figure 3.2 Typical single colony formations (wrinkled, mucoid) of B. licheniformis E30C11 strain on MPC agar.

58

Table 3.7 Summary of the biochemical tests and expected identity of the Bacillus species (Appendix 2) excluding the B. licheniformis isolates already identified by PCR specific gene (Appendix 3).

Isolates Simmons citrate Mannitol NaCl Gelatine Expected agar yolk tolerance test identity polymyxin agar A30C11 + - - + B. subtilis A30C21 + + - + B. cereus B30C12 + - - + B. subtilis B30C21 - + - + B. cereus B30C22 - + - + B. cereus B30C24 - + - + B. cereus B55C01B - - + + B. subtilis C30C11A - + - + B. cereus C30C21 - + - + B. cereus C55C02 - - - + B. subtilis D30C01A - + - + B. cereus D30C01B - + - + B. cereus D30C02 - + - + B. cereus D30C03 + - - + B. subtilis D30C04 - + - + B. cereus D30C12 - - - + B. subtilis E30C03 + + - + B. cereus E30C12 - - - + B. subtilis E30C22 - + - + B. cereus E55C01 - - - + B. subtilis F30C01 - - - + B. subtilis F30C02 - - - + B. subtilis F55C11 - - + + B. subtilis

59

Table 3.8 Results from carbohydrate profile API 50CHB kit after 24 to 48 h of incubation towards Bacillus species from Mozzarella WPC80 samples.

B. B. thuringiensis/B. Test Carbohydrate licheniformis cereus B. subtilis B. pumilus Control (Ng= No growth) GLY Glycerol + - + + ERY Erythritol - - - - DARA D-arabinose - - - - LARA L-arabinose + - + + RIB D-ribose + - + + DXYL D-xylose + - + + LXYL L-xylose - - - - ADO D-adonitol - - - - MDX Methyl-BD-xylopyrannoside - - - - GAL D-galactose + - - - GLU D-glucose + + + + FRU D-fructose + + + + MNE D-mannose + - + + SBE L-sorbose + - - - RHA L-rhamnose + - - - DUL Dulcitol - - - - INO Inositol + - + - MAN D-manitol + - + + SOR D-sorbitol + - + - MDM Methyl αD- annopyranoside - - - - MDG Methyl αD- glucopyranoside + - - + N- NAG acetylglucosamine + + - + AMY Amygdaline + - + + ARB Arbutine + + + + ESC Esculin + + + + SAL Salicin + + + + CEL D-cellobiose + - + + MAL D-maltose + + + + LAC D-lactose - - - - MEL D-melobiose - - + - SAC D-saccharose + - + + TRE D-trehalose + + + + INU Inulin + - - - MLZ D-melezitose - - - - RAF D-rafinose + - + - AMD Amidon + + + - GLYG Glycogene + + + - XLT Xylitol - - - - GEN Gentiobiose - - + - TUR D-turanose + - - + LYX D-lyxose - - - - TAG D-tagatose + - - +

60

DFUC D-fructose - - - - LFUC L-fructose - - - - DARL D-arabitol - - - - LARL L-arabitol - - - - GNT Potassium gluconate + - - -

61

3.3.3 Identification by PCR

The identification of suspect B. licheniformis isolates based on colony morphology was confirmed as expected using species specific PCR based on gyr B gene. A negative control (no template) DNA was included. B. licheniformis was the predominant isolate making up 33 of the 50 isolates (Table 3.9).

3.3.4 Partial 16S rDNA gene sequencing

B. cereus and B. subtilis identified from biochemical tests were subjected to partial 16S rDNA gene sequencing and the results are showed in Table 3.9.

The following isolates (A30C11, B30C12, C55C02, D30C03, F30C01 and F30C02) confirmed as B. licheniformis instead of B. subtilis and E30C21 confirmed as P. glucanolyticus using partial 16s rDNA gene sequencing. All the B. licheniformis isolates identified by PCR and biochemical tests were 16S rDNA sequenced for confirmation.

Table 3.9 The summary of isolates retrieved from WPC80 and their frequency based on specific PCR and 16s rDNA analysis.

No. Bacterial Identity 16S rDNA (% similarity) Frequency 1 Bacillus licheniformis 98-99 33 2 Bacillus cereus/ Bacillus 98 9 thuringiensis 3 Bacillus subtilis 99 2 4 Bacillus pumilus 99 2 5 Paenibacillus glucanolyticus 98 1 6 Lactobacillus plantarum 98 3

Total 50

62

SEM images show B. licheniformis and P. glucanolyticus on membrane surfaces (Figures 3.3 and 3.4 respectively). The spores of P. glucanolyticus have quite a distinctive morphology (Figure 3.5).

Figure 3.3 SEM image of B. licheniformis cell with 3 μm length.

63

Figure 3.4 SEM image of P. glucanolyticus cell with 3 μm length.

64

Figure 3.5 TEM image of P. glucanolyticus spore (Appendix 4).

65

3.4 DISCUSSION

The source of microbiological contamination of Mozzarella WPC80, a quality problem for the dairy industry, has not been thoroughly investigated. The objective of this chapter was to identify the bacteria isolated from Mozzarella WPC80 powder. Six Bacillus species including (percentage of isolates in brackets) Bacillus licheniformis (66%), Bacillus cereus/ Bacillus thuringiensis (18%), Bacillus subtilis (4%), Bacillus pumilus (4%), Lactobacillus plantarum (6%) and Paenibacillus glucanolyticus (2%) were identified based on biochemical and further confirmation were done using molecular analysis in six different WPC80 powder batches (Zain et al, 2016). Mesophilic spore counts from 5 of the 6 WPC80 samples were less than 100 CFU/g and one sample (Sample A) did not yield any colonies on MPCA.

The dominant isolates were B. licheniformis with 33 isolates. These aerobic- spore forming bacteria are frequent contaminants isolated from dairy milk or dairy processing environments (Aouadhi et al, 2013; Buehner et al, 2014; Burgess et al, 2013; Alvarez-Ordóñez et al, 2014; Lücking et al, 2013; Masiello et al, 2014; Frank, 1997), predominant in milk powder (Yuan et al, 2012; Reginensi et al, 2011) and but not commonly reported in whey powder. Buehner et al, (2014) reported that B. licheniformis was the major contaminant in milk regardless of the seasons (winter or summer). B. licheniformis can contribute to product spoilage through the production of protease enzymes (Teh et al, 2012).

Infrequently, Bacillus species such as B. licheniformis, B. amyloliquefaciens and B. pumilus have been reported to produce toxic components which may play role in food poisoning (Lücking et al, 2013). The origin of B. licheniformis is believed to be from soil and the heat resistant spores survive pasteurisation of milk and further survive the cheese manufacturing process thus probability contaminating the cheese product too. Although the isolates were recovered at 30°C (), some grew at 55°C (thermophilic)(Lücking et al, 2013). Watterson et al, (2014) reported that sweet whey and non-fat dry milk showed higher prevalence of thermophilic and mesophilic spore-formers compared with acid whey and WPC80. Three of the B.

66

licheniformis isolated in the present trial grew and formed biofilm at both 30°C and 55°C. Facultative thermophiles in dairy processing belong to the Bacillus genus and include species such as B. coagulans, B. licheniformis and B. pumilus that are able to grow at both mesophilic and thermophilic temperatures (Burgess et al, 2013; Flint et al, 2011a; Flint et al 2011b).

The second most isolated Bacillus species in the present study was B. cereus/ B. thuringiensis with 9 isolates. B. cereus is capable of producing emetic and diarrhoeal toxin that can cause to food poisoning (Burgess et al, 2013; De Jonghe et al, 2010). The next most isolated species were B. subtilis and B. pumilus. These three Bacillus species are also common contaminants in dairy products (Masiello et al, 2014; Pasvolsky et al, 2014). Certain mesophilic Bacillus species. (e.g., B. subtilis, B. pumilus, and B. licheniformis) can also grow at temperatures that are used to detect and enumerate thermophilic organisms (i.e., 55°C) (Yuan et al, 2012). In one recent trial, WPC80 contained thermophilic spore counts detected by direct plating (Watterson et al, 2014). Three Lactobacillus plantarum isolates were also isolated from the Mozzarella WPC80 samples. This is not the first time that Lactobacillus species have been associated with whey as Tang et al, (2011) isolated Lactobacillus species from the permeate side of UF membrane used in a whey processing plant.

Based on the 16s rDNA gene sequence analysis (Table 3.6), one isolate was identified as P. glucanolyticus. This is the first time that this bacterium has been isolated from Mozzarella WPC80 and there are no records of this bacterium being isolated in New Zealand. It was identified using 16S rDNA sequence with confident level 99% (BLAST) and 94% (Seqmatch) and on the basis of a biochemical profile obtained by the use of API 50CHB (bioMérieux). The result was two possible identities for the P. glucanolyticus. There was a 78.7% similarity to P. amylolyticus and 19.0% similarity to P. glucanolyticus based on the API 50CHB test, demonstrating the difference between biochemical and molecular typing methods.

The P. glucanolyticus was observed using SEM (Figure 3.4) and the intercellular properties of the cells together with the spores were observed using TEM by Manawatu Microscopy Imaging Centre. The cells were long (> 3.0 mm and thin (<

67

0.9 mm) and produced oval terminal spores that markedly distended the sporangium. Interestingly, their spores portrayed a very unique shape, different from spores of Bacillus species (Figure 3.5).

This bacterium formerly included in the genus Bacillus is a facultative anaerobic, long, thin rod-shaped bacterium with terminal spore formation and may be isolated from various soils. Paenibacillus species are psychrotrophic spore-forming bacteria that have been isolated from farm environments and raw and pasteurised milk and have the potential to cause spoilage (Lorentz et al, 2006; Ferrand et al, 2013). Lorentz et al, (2006) studied the antimicrobial activity by Paenibacillus species against bacteria, filamentous fungi and yeast isolates with the purpose of finding new bacteria for microbiological control. His findings noted that P. glucanolyticus produced antibiotic peptides that inhibited Xanthomonas axonopodis growth with inhibition zones from 12 to 18 mm. Ferrand et al, (2013) published a case report on P. glucanolyticus infection in a 65-year old patient with type 2 diabetes who developed a cardiac device-related endocarditis. This is believed to be the first ever report associating P. glucanolyticus with human infection.

68

3.5 CONCLUSIONS

Mesophilic bacteria, in particular, B. licheniformis, may predominate in Mozzarella WPC80 contributing to aerobic plate counts close to or exceeding specification limits. Preliminary attempts to identify the contaminants using biochemical tests gave an indication to bacterial identity and later this was confirmed by bio-molecular tests. Both biochemical and bio-molecular were in agreement in this study. One bacterium, P. glucanolyticus which recorded as the first Paenibacillus species isolated from Mozzarella WPC80 in New Zealand should be highlighted as it is a psychrophilic, spore forming bacteria that can potentially thrive during cold UF process. Dairy practice should include a lower incubation temperature than 30°C to enable isolates of this bacterium to be detected from routine inspection of their finished product. However, B. licheniformis is the predominant isolate and this cannot thrive in the temperatures used for UF process. The growth of B. licheniformis in other parts of a whey manufacturing plant will be the focus of this study.

69

______Chapter 4 ______

ATTACHMENT AND BIOFILM FORMATION BY B. licheniformis

70

4.1 INTRODUCTION

Contamination of dairy plants by Bacillus species can cause product quality problems as their spores cannot be destroyed by heat treatment such as pasteurisation or thermalisation (De Jonghe et al, 2010). Bacteria contaminating dairy manufacturing plant are often growing as a biofilm (Flint, 1998; Flint et al, 2000; Tang et al, 2009). Initially, UF process was thought to be the most likely source of biofilm and product contamination in a WPC manufacturing plant however the results identifying the predominant contaminant from Chapter 3, suggest this is unlikely. Other surfaces in the WPC manufacturing plant will now be considered as the site for biofilm growth. These will include heat exchangers (SS surface) used to treat the whey before WPC manufacture (thermalisation), the preheaters for the concentrated whey before evaporation and the evaporators used to concentrate the whey before drying.

The conditions at these points in the manufacturing process such as temperature, water activity, whey protein, salt composition and lactose content influence biofilm formation and will determine the most likely zone for biofilm growth. All these factors will vary during the process and are likely to have an influence on biofilm development. For example, Somerton et al, (2013) found that changes in salt concentration influenced the attachment of Geobacillus and Anoxybacillus species to SS surfaces. Understanding the source of contamination and development of B. licheniformis biofilms will be important in developing methods to control or prevent biofilm formation and the contamination of WPC. This chapter aimed to investigate the attachment and biofilm formation abilities by B. licheniformis isolated from Mozzarella WPC80 samples.

71

4.2. MATERIALS AND METHODS

4.2.1 Source of strains

Thirty three B. licheniformis strains isolated from WPC80 samples as retrieved from WPC80 samples in Chapter 3, were used in this study. The isolates were maintained on nutrient agar slants at 4oC and in glycerol beads stock at -80oC.

4.2.2 Attachment and biofilm screening

A microtitre plate assay (Oh et al, 2007) was used to determine the ability of isolates to attach (adhere to a surface) and grow (reproduce on a surface) and form biofilm based on crystal violet (CV) absorbance. A sterile 96-well flat-bottomed polystyrene microtitre plate (Falcon® 96, 35 3072, Becton, Dickinson & Company, USA) was filled with 230 μL of TSB. Three wells were filled with TSB only as negative controls. Overnight culture (20 μL) was added into the test wells, three wells per culture, and the plates were incubated aerobically for 1 h for attachment studies at 30°C and 55°C to reflect the temperatures used to isolate them from WPC80 powder. A time of 1 h was selected to ensure irreversible attachment of the cells to the surface with minimal opportunity for growth.

For the biofilm study, overnight culture (20 μL) was added into the test wells, three wells per culture, and the plates were incubated 24 h at 30°C, 37°C and 55°C. This reflected the temperatures used to isolate these bacteria but also the best growth temperature for B. licheniformis and to reflect temperatures in different parts of the manufacturing plant. A time of 24 h was considered sufficient to allow biofilm to develop.

After incubation, the contents of the microtitre plates were removed by inverting the plates, and then the wells were washed three times with 300 μL of sterile distilled water. The remaining attached cells were fixed with 250 μL of pure methanol per well for 15 minutes. The liquid was discarded and the microtitre plates air-dried. The microtitre plate wells were stained with 250 μL of 0.5% (w/v) CV for five minutes. The excess stain was rinsed off by placing the microtitre plate under running

72

distilled water. After the microtitre plates were air-dried, the dye bound to the adherent cells was re-solubilised with 250 μL of 33% (v/v) glacial acetic acid per well. The optical density (OD) of each well was measured at 570 nm using an automatic 96-well microplate reader (BMG Labtech Spectrostar microplate reader, Bio-Tek Instruments, INC, Winooski, VT, USA). The biofilm formation of each B. licheniformis isolate (the predominant isolate) was tested using the microtitre plate assay with the strength of biofilm formation based on the following range of OD values based on (Oh et al, 2007). The cut-off OD (ODc) was defined as three SDs above the mean OD of the negative control. Strains were classified as follows: OD < ODc = no biofilm former, ODc < OD < (2 x ODc) = moderate biofilm former and OD > (2 x ODc) = strong biofilm former.

4.2.3 Biofilm formation on SS using three different concentrations of reconstituted whey

Reconstituted whey (RWPC80) medium and RWPC80 with lactose and minerals (artificial whey permeates) were prepared in concentrations of 1%, 5% and 20% (pH 5.7 - 6.5) as described in Tang, (2011) to copy the composition of whey protein concentration in three different stages of UF process. The minerals of artificial whey permeate listed in Table 4. 1, were prepared by mixing the listed minerals in deionised water to make up 1 L (pH 6.0 - 6.5) and sterilised. WPC80 powder was sterilised by gamma irradiation (25 kGy, MSD, Upper Hutt, Wellington, New Zealand) prior to medium preparation to avoid any competition from the natural microflora of whey. An initial inoculum of 6 log10 CFU/mL of one B. licheniformis isolate (E30C11) was cultured in each RWPC80 concentration. SS is the most common material used in food processing plant surfaces as it is corrosion resistant and easy to clean as described by Holah & Gibson, (2000). A 304 grade SS coupon (1 cm2) was used as the surface for this B. licheniformis biofilm study. Each RWPC80 solution contained one SS coupon (1 cm x 1 cm) submerged into the medium and incubated at 37°C for 24 h. The SS coupon was aseptically removed from the medium and rinsed twice using sterilised deionised water and the cells that had colonised the surface were

73

removed by bead beating using 3g of sterile glass beads (d= 2 mm) by vortex mixing at high speed for 2 minutes. The detached cells were serially diluted to 10-3 and MPCA was used for pour plating in duplicate and incubated at 37°C for 24 to 48 h. Colonies forming on MPCA were counted. Mean and standard errors were taken from duplicates.

Table 4.1 The mineral content of artificial whey permeate.

Mineral Amount Source KOH 52.7 mL 2 mol per litre BDH, Poole, England C H O Na .2H O 24.29 g Merck KGaA, Darmstadt, 6 5 7 3 2 (Trisodium citrate Germany dehydrate) C H O K .2H O 4.99 g UNIVAR, Auckland, NZ 6 5 7 3 2 (Tripotassium citrate dehydrate)

CaCl .2H O 3.67 g Biolab, Clayton, Australia 2 2 (Calcium chloride)

MgCl .6H O 5.85 g J.T. Baker, Phillipsburg, 2 2 (Magnesium chloride) Mexico

KH PO 23.3 g Merck KGaA, Darmstadt, 2 4 Germany

H SO 17.1 mL 3 mol per litre Biolab, Clayton, Australia 2 4

4.2.4 Biofilm formation on SS using three different media

An initial inoculum of 6 log10 CFU/mL of B. licheniformis (diluted and confirmed with plate counts) was grown in three different media; tryptic soy broth (artificial media), 1% RWPC80 and 1% RWPC80 with additional lactose and minerals (artificial whey permeates) to mimic the composition of whey protein concentration

74

in the first stage of UF. Each of the three media preparations contained one SS coupon and was treated as in section 4.2.3.

4.2.5 Biofilm formation on SS with effects of individual cations

One B. licheniformis isolate derived from each WPC80 powder samples (A30C11, B55C11, C55C02, D55C03, E30C11 and F30C02) were chosen to study on the effects of individual cations; calcium (Ca2+) and magnesium (Mg2+). Each of these stock cations was prepared by dissolving 3.67g Calcium chloride (CaCl2.2H2O) in 1

L of sterile deionised water; 5.85 g Magnesium chloride (MgCl2.6H2O) in 1 L of sterile deionised water. These individual cations stocks were used to replace the artificial whey permeate (used previously) in 1% RWPC80 with lactose media for biofilm study on effects of these individual cations. The concentrations of individual cations were 0.025 mol per L for Ca2+ and 0.029 mol per L for Mg2+ and substituted the whey permeates as 6% from the 1% RWPC80 medium to replicate the concentration of whey medium in the formulation. The amount of each cation used were 200 mg/L. In real scenario, the value of minerals such calcium and magnesium are 423 mg/100g and 50 mg/100g respectively (Appendix 5).

4.2.6 Statistical analysis

Statistical calculation was performed on the log density values with the mean and standard deviations produced from the log density of 3 replicates. Anova one- way (Excel, 2016) was used to analyse the variance of individual cations affecting biofilm formation of B. licheniformis isolate on SS.

75

4.3 RESULTS

4.3.1 Attachment and biofilm study of B. licheniformis isolates

The attachment of a subset of mesophilic thermotolerant B. licheniformis isolates was determined using a microtiter plate with TSB as the growth medium. Successful attachment was based on an OD that exceed 3 x SD of negative control (TSB). Two temperatures were selected for this preliminary attachment study; 30°C and 55°C. Figure 4.1 showed 3 of 11 B. licheniformis did not attach on the microtitre plate surfaces at 30°C after 1 h of incubation as the reading was below the negative control. Ten selected B. licheniformis isolates were able to attach on the microtitre plate surface at 55°C after 1 h of incubation in a static condition (Figure 4.2).

The biofilm formation of each B. licheniformis isolate was tested using the microtitre plate assay with the strength of biofilm formation based on the following range of OD values (Table 4.2) (Oh et al, 2007). In this study, three temperatures were used; 30°C, 37°C and 55°C on the basis of the fact that B. licheniformis as it is considered as mesophilic/thermo-tolerant bacteria. The cut-off point (ODc) value at 570 nm was 0.5.

The biofilm formation by 33 B. licheniformis and the amount of biofilm at three different temperatures are given in Figure 4.3 and Table 4.3 respectively.

76

0.1

0.09

0.08

0.07

0.06

0.05

0.04 1h Absorbanceat 570 nm 0.03

0.02

0.01

0

Mesophilic B. lichenformis

Figure 4.1 B. licheniformis attachment on a microtitre plate surface after 1 h of incubation at 30°C. Values and error bars represent the mean and standard deviation respectively for 8 independent replicates.

77

0.06

0.05

0.04

0.03

1h 0.02 Absorbanceat 570 nm

0.01

0

Thermotolerant B. licheniformis

Figure 4.2 B. licheniformis attachment on microtitre plate surface after 1 h of incubation at 55°C. Values and error bars represent the mean and standard deviation respectively for 8 independent replicates.

78

4

3.5

3

2.5

2 30C

1.5 37C 55C Optical Density at 570nm 1

0.5

0

B. licheniformis isolates

Figure 4.3 Biofilm formation by 33 B. licheniformis isolates in TSB at three different temperature profiles. Values and error bars represent the mean and standard deviation respectively for 20 independent replicates.

79

Table 4.2 Biofilm categories determined by OD absorbances at 570 nm.

Range Indication X ≤ 0.5 No biofilm 0.5 ≤ X ≤ 1.0 Moderate biofilm X ≥ 1.0 Strong biofilm

Table 4.3 Summary of frequency of B. licheniformis biofilm categories at three different temperatures.

Tryptic Soy Broth & Biofilm formation Temperature None Moderate Strong

30°C 30 0 3

37°C 21 3 9

55°C 30 2 1

80

4

RWPC80

3

RWPC80 with

-2 lactose and minerals 2 Log CFU cm LogCFU

1

** **

0 1520 Media concentration (%)

Figure 4.4 The 1% reconstituted WPC80 with and without lactose and minerals was the best medium for B. licheniformis E30C11 isolate to form biofilm on SS after 24 h incubation at 37°C compared to 5% and 20% RWPC80. The ** in the graph indicates no growth/biofilm formation. Values and error bars represent the mean and standard deviation respectively for 3 independent replicates.

81

6 TSB

5 1% RWPC80

1% RWPC80 4 with lactose -2 and minerals

CFU cm 3 10 Log

2

1

0

B. licheniformis isolates

Figure 4.5 Biofilm formations by 33 B. licheniformis isolates on SS submerged in three different mediums at 37°C for 24 h. These 33 B. licheniformis isolates are from six Mozzarella WPC80 powder abbreviated A, B, C, D, E and F. Values and error bars represent the mean and standard deviation respectively for triplicate assays (N=3).

82

Prior to using SS as surface for biofilm study, biofilm formation by 33 B. licheniformis in WPC was assayed using the microtitre plate assay at 37°C but the CV absorbance readings were high (3.5 and above) at 570 nm (results not included). Protein precipitation was observed on the bottom and side of the microtitre plate wells after 24 h of incubation in 1% RWPC80 and 1% RWPC80 with lactose and minerals.

Three different RWPC80 concentrations (1%, 5% and 20%) were tested for B. licheniformis biofilm formation (Figure 4.4). The 1% RWPC80 was the best (with and without lactose and minerals) medium for B. licheniformis E30C11 to form biofilm on SS after 24 h incubation at 37°C.

Nine isolates that formed strong biofilm in the microtitre plate assay using TSB -2 medium at 37°C showed strong biofilm on SS (log10 CFU cm ) (Figure 4.5). Three isolates (C55C02, C55C11 and F30C14) that formed moderate biofilm in the microtitre plate assay in TSB, produced strong biofilm on SS in TSB but not in 1% WPC80 medium. Twenty one isolates that did not produce biofilm in the microtitre plate assay in TSB medium at 37°C, showed strong biofilm formation on SS in TSB medium. Seventeen isolates showed better biofilm with the enrichment of lactose and minerals. Four isolates from product batch F (Chapter 3) showed biofilm formation in 1% RWPC80 with the enrichment of lactose and minerals but none in 1% RWPC80 without minerals.

83

6

* 5 * * * 4 * -2 * * 3 * Log CFU cm LogCFU 2

1

0 A30C11 B55C11 C55C02 D55C03 E30C11 F30C02

Isolates

Figure 4.6 Biofilm formation by six B. licheniformis isolates from each isolate on SS in 1% RWPC80 with individual cations (Ca2+ and Mg2+) for 24 h at 37oC. Values and error bars represent the mean and standard deviation respectively for triplicate samples in two separate runs. The asterisk (*) represent the significant differences at p < 0.05.

Grey bar: 1% reconstituted WPC80 without added minerals

Stripe bar: 1% reconstituted WPC80 with calcium

Black bar: 1% reconstituted WPC80 with magnesium

84

The individual effects of Ca2+ and Mg2+ were examined on 6 B. licheniformis isolates and there were significant different (p < 0.05) by one-way ANOVA between these two ions on tested strains and their effect on B. licheniformis biofilm formation on SS at 37°C. Magnesium resulted in the most biofilm formation of A30C11, B55C11, C55C02 and D55C03 while calcium resulted in the most biofilm of E30C11 and F30C02 isolates (Figure 4. 6).

85

4.4 DISCUSSION

There is a lack of information on the biofilm forming ability of B. licheniformis and the importance of biofilms of this bacterium in the contamination of dairy products. B. licheniformis are thermo-tolerant bacteria with the ability to attach and form biofilm. This biofilm is a potential source of microbial contamination of product, resulting in microbial specification limits being exceeded. Spoilage of product may result from the metabolic products, such as enzymes, produced by these biofilms. These bacteria are therefore likely to form biofilm on manufacturing plant surfaces where they can grow and then be released into the whey being processed.

Three B. licheniformis isolates (C55C01, E30C11 and F55C12) produced strong biofilm OD570nm ≥ 0.5 at 30°C with an O.D at 3.5 (the highest absorbance from microtitre plate reader) (Table 4.2). Nine B. licheniformis isolates produced strong biofilms at 37°C suggesting 37°C is the best temperature for B. licheniformis to form biofilm (Figure 4.3).

B. licheniformis is known as thermotolerant bacteria and able to grow at temperatures up to 55°C (Dhakal, 2013; Smithers, 2008). However, in this study, most of the B. licheniformis isolated from 6 different whey samples did not produce biofilm at 55°C. Biofilm formation at 55°C was generally poor with only three isolates showing a tendency to form biofilm at this temperature on microtitre plate surface. High temperature will cause high solubility although calcium showed less solubility compared to the minerals (Bansal & Chen, 2006). Calcium phosphate is the main mineral deposited at 70% - 80% on SS at temperature above 110°C. Moderate biofilm formation in TSB at 37°C was produced by three B. licheniformis isolates. The majority of B. licheniformis strains were unable to produce biofilm at any of the three temperatures (30°C, 37°C and 55°C) in the TSB medium. However, they may produce biofilm under the conditions of a dairy manufacturing plant on a protein fouled surface. Most of the facultative anaerobic and spore forming B. licheniformis were observed to produce pellicles after 24 h incubation at 30°C, 37°C and 55°C in TSB medium.

86

B. licheniformis (E30C11) was able to form biofilm in a microtitre plate assay using tryptic soy broth medium at 30oC, 37oC and 55oC. This strain was selected from 33 B. licheniformis isolates to determine the optimum concentration of whey to promote biofilm formation (Figure 4.4). A preliminary study using microtitre plates showed protein precipitation on the microtitre plate surfaces when using RWPC80 and RWPC80 with lactose and minerals as the substrate, hence giving high CV readings at 570 nm (results not included) that reflected the protein precipitation rather than the microbial colonisation. SS was chosen as the surface to study biofilm formation as it is a common surface material used in dairy manufacturing plant. E30C11 produced -2 1.82 log10 CFU cm in 1% RWPC80 while no biofilm formation was observed at 5% and 20% of RWPC80 at 37oC for 24 h in a static environment. On the other hand, E30C11 showed more biofilm in 1% RWPC80 with lactose and minerals (artificial -2 whey permeates) with 3.29 log10 CFU cm . There was biofilm formation observed in -2 5% and 20% RWPC80 with lactose and minerals with 2.31 log10 CFU cm and 1.53 -2 log10 CFU cm respectively (Figure 4.4).

Thermo-tolerant B. licheniformis can survive pasteurisation. They can grow in subsequent manufacturing steps on equipment such as cream separators, heat exchangers, preheaters and evaporators during milk powder production (Dhakal, 2013). The plate heat exchanger temperature profile is presented in Appendix 1. The temperature ranges between 37°C to 77°C with cold and heated whey in the regenerative zones of the plate heat exchanger. Scott et al, (2007) and Murphy et al, (1999) concluded that spore forming bacteria contamination occurred in evaporators after a 20 h run time. Based on the results in this present study (Figure 4.4), B. licheniformis contamination most likely occurs before UF as biofilm counts were highest at 1% RWPC80 and the role of lactose and minerals, removed during dialysis in UF will have an impact on B. licheniformis biofilm on SS. The water activity at pre- UF sites is expected to have minimal effect on B. licheniformis biofilm growth however, after evaporation the lower water activity and higher total solids is expected to limit microbial growth. Figure 4.4 showed E30C11 strain able to form biofilm in 1%, 5% and 20% RWPC80 medium with lactose and minerals.

87

As 1% RWPC80 was favoured by E30C11 to form biofilm, further study was done using the 33 B. licheniformis isolates with three different media; 1% RWPC80, 1% RWPC80 with lactose and minerals and artificial medium (TSB). There were variations in biofilm formation by B. licheniformis on SS with each medium (Figure 4. 5). Twenty one B. licheniformis were non-biofilm formers using TSB as medium at 37oC using the microtitre plate assay. However, only one B. licheniformis isolate (F30C02) from 33 B. licheniformis did not form biofilm on SS whereby the rest -2 produced between 1.81 – 4.63 log10 CFU cm in TSB. Biofilm formation by 28 B. -2 licheniformis on SS resulted in 1.38 to 3.47 log10 CFU cm using 1% RWPC80 as media. Interestingly, five isolates from the same batch of WPC (B55C11, F30C02, F30C11, F30C14 and F55C02) did not produce any biofilm in 1% RWPC80 but -2 produced between 2.46 to 3.87 log10 CFU cm in 1% RWPC80 with lactose and minerals which indicates the importance of lactose and salts such as calcium and magnesium for these strains to form biofilm in a whey environment. The other B. -2 licheniformis isolates showed increased biofilm count (1.6 to 5.07 log10 CFU cm ) with 1% RWPC80 with lactose and minerals substrates except for D55C04 which did not produce biofilm. Dhakal et al, (2013) reported that B. licheniformis occurs in milk and dairy products as many different genotypes that vary in the term of heat resistance and biofilm capabilities. This could explain what was seen in the present investigation.

Subsequent tests for biofilm formation on SS showed an increased frequency of biofilm formation with 32/33 strains forming biofilm in TSB at 37°C. This demonstrates the limitation of the microtitre plate assay for screening for biofilm formation. This suggests that biofilm growth of B. licheniformis favours a SS surface. The results of biofilm formation on SS coupons in this study are expected to differ from the actual plate heat exchanger in the industry as the experiments were conducted in batch not continuous condition. However, the role of B. licheniformis biofilm formation on SS as source of product contamination cannot be ruled out.

Ions have been reported as having a role in biofilm formation. Bellona & Drewes, (2005) reported that positively charged ions such as sodium, calcium, magnesium, and cationic surfactants can bind and neutralise negatively charged

88

surfaces and enhance bacterial attachment to a filter membrane. The role of precipitated minerals in biofilm development in WPC80 production is aiding the conditioning film formation (Zaky et al, 2012; Tang et al, 2015; Teh et al, 2015). Somerton et al, (2013) found that changes in salt concentration (calcium and magnesium) influenced the attachment of Geobacillus and Anoxybacillus species to SS surfaces. It has either positive (enhance) or negative (inhibit) biofilm development depending on the strains and concentration (Somerton et al, 2015). In this chapter, 1% RWPC80 with additional lactose and minerals (calcium and magnesium) did increase the magnitude of biofilm formation by B. licheniformis isolates by 1-2 fold except for one strain (D55C04) which was unable to form biofilm. Individual cations (Ca2+ and Mg2+) at 200 mg/L in these experiments were significantly (p < 0.05) affecting B. licheniformis biofilm formation on SS with Mg2+ increasing the biofilm of 4 of the 6 B. licheniformis biofilm on SS. High minerals needed for the biofilm formation. Traces of minerals (calcium and magnesium) were recorded 423 mg per 100g and 50 mg per 100g (Appendix 5). In this study, both minerals were added at 200 mg/L using 1g of sterile WPC80 powder. On the other hand, Oknin et al, (2015) reported that magnesium ions mitigate biofilm formation by Bacillus species. They studied the effects of Mg2+ on biofilm formation of B. subtilis in chemically defined medium in flasks. The biofilm formation was quantified by the production of pellicles on the surface of the medium and images under confocal laser microscopy. This is a general result as different bacteria and methods were used, and different surfaces were compared in the present study.

89

4.5 CONCLUSIONS

B. licheniformis is a contaminant of WPC. Biofilm formation by B. licheniformis isolates from WPC varies between strains and is dependent on the substrate surface with SS supporting more growth than polystyrene. There were limitations on microtitre plate assay to study biofilm formation on polystyrene where protein precipitation was developed instead of bacterial biomass and hydrophobicity interaction cannot be ruled out.

Whey concentration of 1% RWPC80 with and without lactose and minerals supports biofilm formation of B. licheniformis with most biofilm forming in the presence of lactose and minerals. The ability of B. licheniformis to form biofilm has been demonstrated and this is important as this is the most likely source of contamination of high numbers of B. licheniformis and their enzymes in product.

90

______CHAPTER 5 ______

CHARACTERISATION OF B. licheniformis

91

5.1 INTRODUCTION

B. licheniformis biofilm growth was tested at 30oC, 37oC and 55°C in Chapter 4. The origins of Mozzarella WPC80 powders were from cold UF (10°C) concentration and dialysis, followed by evaporation and spray drying. The hypothesis in Chapter 1.3 was that the contaminants formed biofilms on UF membrane as it represents the largest surface area in the WPC plant. Therefore, in this chapter, growth at 10°C was tested.

Lactose is a major carbohydrate in whey making up approximately 5% of the total of cheese whey. It is postulated that microorganisms that contaminate whey use lactose as their carbohydrate source using β-galactosidase to break down the lactose into fermentable sugars (Geiger et al, 2016; Juajun et al, 2011). B. licheniformis is one of the microorganisms commonly found in whey and is regarded as a potential cause of spoilage through the production of heat stable enzymes such as protease and lipase. Microbial spoilage enzymes produced by bacteria may result in product defects in terms of organoleptic and functionality changes. There is some concern that B. licheniformis may be involved in nitrate reduction resulting in high nitrite levels in whey products. The importance of enzyme production and nitrate reduction due to B. licheniformis in whey is unknown. The percentage of isolates of B. licheniformis from whey, producing protease or lipase and reducing nitrate that may affect the quality and safety of whey products respectively is unknown.

The bacterial contaminants, such as B. licheniformis, in many dairy products are believed to originate from biofilms on dairy manufacturing plant surfaces. These dairy biofilms contain extracellular polymeric substances (EPS) and milk residues such as protein and calcium phosphate (Flint, 1998). The EPS is an integral part of a biofilm. EPS consists of polysaccharides, eDNA, proteins and lipids played vital role in the adhesion and aggregation of bacterial cells. In addition, EPS form a protective barrier and provide a nutrient source for the biofilm communities as energy storage (Simões et al, 2010).

Dhakal, (2013) studied the heat resistance of B. licheniformis spore in milk and Husmark, (1993) investigated the high temperature (80°C) effects towards B.

92

licheniformis spore hydrophobicity. However, the heat resistance of B. licheniformis vegetative cells has not been reported. It is important to understand the heat resistance of B. licheniformis vegetative cells towards high temperature such as pasteurisation in dairy processing.

This chapter aimed to characterise 33 B. licheniformis isolates from whey product manufacture by studying their heat resistance, ability to ferment lactose, produce enzymes, convert nitrate to nitrite and produce EPS and biosurfactant that will contribute to their survival through dairy manufacture and determine their influence on product quality.

5.2 MATERIALS AND METHODS

5.2.1 Growth at 10°C

Twenty μL of 33 B. licheniformis isolates were inoculated in 230 μL TSB (pH 7.0 ± 0.2) in 96-well microtitre plate wells in 8 replicates and incubated at 10°C for 24 h. Three wells of microtitre plate were filled with sterile TSB as negative controls. Growth was measured at 570 nm using a microplate reader (Spectrostar Nano, BMG Labtech, Auckland, New Zealand). Similar growth trials were done in 1% RWPC80 with lactose and minerals, except cultures were prepared in 25 mL bottles (2 mL inoculum in 23 mL medium) instead of the microtitre plates and incubated at 10°C for 24 h. B. licheniformis growth was determined by serial dilution and colony formation on MPCA after incubation for 24 at 10°C. The purpose of these trials was simply to indicate the potential for growth at 10°C in either medium rather than do a direct comparison of the extent of growth.

93

5.2.2 Lactose fermentation

In order to determine the ability of B. licheniformis isolates to ferment lactose, phenol red broth was prepared following the instructions of the manufacturer (Merck, BDH, Palmerston North, New Zealand), distributed into test tubes and autoclaved. Lactose 10% m/v stock solution was prepared by adding 10.00 ± 0.001 g of lactose monohydrate into 100 mL beaker and made up the volume by adding sterile deionised water and mixed well. The lactose stock solution was sterilised by filtration using a 0.2 μm membrane filter and stored at 4°C before use. An amount of 1 mL of lactose stock solution was pipetted into a test tube containing 9 mL of sterile phenol red broth to make a 1% lactose solution. Lactose fermentation is indicated by the change in phenol red from red to yellow, indicating acid production.

5.2.3 Protease and lipase enzyme

To determine the ability of 33 B. licheniformis to produce protease, a single colony was streaked onto calcium caseinate agar (Composition: meat peptone, sodium chloride, beef extract, casein, calcium hydroxide, calcium chloride, bacteriological agar) (Conda, Spain). The plates were incubated at 37°C and checked daily for 2 days for the presence of clear zones along the streaked line.

To determine the ability of 33 B. licheniformis isolates strain to produce lipase, a single colony was streaked onto Tributyrin agar (Fort Richard Laboratories Ltd, NZ) and spirit blue agar (Composition: Pancreatic digest agar of casein, yeast extract, agar, spirit blue) (DifcoTM, Becton Dickinson, USA) with soy oil addition. The plates were checked between 24 to 48 h for the presence of clear zones along the streaked lines.

94

5.2.4 Nitrate conversion

B. licheniformis isolates were incubated in nitrate broth (9 g/L) at 37°C for 24 h. One drop of sulfanilic acid and one drop of a α-naphthylamine were added into each broth. A colour change from yellow (nitrate broth) to red as a result from the reaction between the added two reagents, indicates nitrate reduction. If the yellow colour of nitrate broth remains after the addition of two reagents, half teaspoon of zinc was added into the broth. This step was a confirmatory test which zinc is able to catalyse the reduction of nitrate to nitrite. No colour change means B. licheniformis reduced nitrate completely to ammonia and nitrogen gas. Therefore, no colour change at this point is a positive result. If the broth colour changed to red, it indicated negative results as the nitrate was reduced to nitrite by zinc and not by B. licheniformis.

5.2.5 Haemolysis on Columbia sheep blood agar

To indicate the production of lichenysin, haemolysis of sheep blood is used (Madslien et al, 2013). B. licheniformis isolates were streaked on to Columbia blood sheep agar plates with 5% sheep blood (Fort Richard Laboratories Ltd, NZ) and incubated at 37°C for 24 to 48 h. The presence or absence of clear zone surrounding the colonies (haemolysis) was recorded after 24 to 48 h.

5.2.6 Lichenysin synthetase gene (Lch AA)

Sheep blood haemolysis is an indicator of lichenysin production. However, haemolysis could also be due to other factors. To provide further evidence for the ability to produce lichenysin, a PCR test was used to determine whether the B. licheniformis isolates contained the Lch AA gene that is one of the genes that are responsible for the secretion of lichenysin. Lichenysin synthetase A (Lch AA) was indicated by PCR amplification using the following primers;

95

F: 5’ –ACTGAAGCGATTCGCAAGTT- 3’

R- 5’- TCGCTTCATATTGTGCGTTC- 3’

The PCR conditions were; five minutes of denaturation at 95°C, 35 cycles denaturation at 95°C for 10s, annealing at 56°C for 10s, extension at 72°C for 30s and elongation step at 72°C for 7 minutes (Madslien et al, 2013). The PCR products were visualised (E-Gel iBASE™, Invitrogen) using pre-made 2% agarose electrophoresis gel (E-Gel ® EX with SYBR Glod II) then visualised under UV Transilluminator, UVP, Inc (Chromator-Vue, San Gabriel, Californis, USA) and UVITEC (Cambridge, UK). A negative control (no template) was used.

5.2.7 Pellicle formation and Congo red binding assay

Pellicle formation and Congo red binding were used to indicate EPS production. The production of pellicles indicates the overproduction of polysaccharides by B. licheniformis and is indicative of biofilm formation. Pellicle formation was studied by inoculating 1 loop of B. licheniformis single colony in 10 mL sterile TSB in a 25 mL universal bottle (25 mm by 70 mm). The bottles were incubated in a static condition at 37oC for 2 days. Pellicle formation at the air-liquid interface was visually inspected and the images were captured using a digital camera (Nikon, Japan).

To quantify the amount of polysaccharide produced by B. licheniformis, a Congo red binding assay was used. This method was adapted and modified from Ghafoor et al, (2011) and Spiers et al, (2003). Each of the 33 B. licheniformis isolates was inoculated into 10 mL TSB broth for 48 h at 37oC in a static condition. The bacterial mass and polysaccharides of B. licheniformis isolate cultures were collected by centrifugation and the supernatant was discarded. The pellet was washed by using sterile deionised water and transferred into 2 mL microfuge tubes. The pellet was re- suspended in 1 mL of 40 μg/mL Congo red in sterile deionised water and incubated for 90 minutes with shaking (37°C at 100 rpm). Bacterial mass and bound Congo red

96

was sedimented by centrifugation at 13 500 x g for 6 minutes. The supernatant was collected and the amount of Congo red remaining in the supernatant was determined by measuring the OD 490nm of the solution and the amount of EPS was calculated based on the formula below to obtain the percentage of polysaccharides produce by B. licheniformis strains.

OD 490 control - OD490 final X 100%

OD 490 control

5.2.8 Heat resistance study at 72oC, 75oC and 80°C

A single colony of selected B. licheniformis isolates was isolated into 20 mL TSB and incubated for 24 h at 37°C. The B. licheniformis cells were observed using Gram staining under light microscopy to confirm the vegetative cells and no spore production prior the heat kinetic study. An immersed coil apparatus (Coil 100, Sherwood Instruments, MA, USA) was used to heat a suspension of B. licheniformis culture under selected controlled temperature (70°C, 72°C and 80°C). At the beginning, the water bath attached to the coil apparatus was set to the selected temperature before use. The heating time was set to the required time intervals; 15, 30 and 60 s.

The apparatus was cleaned by injecting 20 mL of NaOH, followed by 50 mL of sterile water, then 20 mL of H2SO4 and 50 mL of sterile distilled water.

Ten μL of the overnight B. licheniformis culture approximately 8.0 ± 0.5 log CFU/mL (confirmed by plate counting on MPCA) was injected into the Coil 100. At the pre-set temperature and time interval, a small sample of heated sample ≈ 1 μL was discharged from the discharge tube into a sterile sample tube. The sample tubes were covered immediately with lids and subjected to serial dilution up until 10-6. One mL each dilution was plated in triplicate on MPCA and incubated at 37°C for 24 to 48 h.

97

5.3 RESULTS

5.3.1 Lactose fermentation, protease, lipase, nitrate conversion, Columbia sheep blood agar haemolysis and lichenysin synthetase gene A

The results for B. licheniformis growth at 10°C, lactose fermentation, protease and lipase enzymes production, nitrate conversion, haemolysis on Columbia sheep blood agar and synthetase A (lch AA gene) analysis are shown in Table 5.1. These results were compared to Bergey’s Manual (Logan & Von, 2015).

98

Table 5.1 The characterisation of B. licheniformis isolates from WPC80 samples

Isolates Growth Lactose Protease Lipasec Nitrite EPS (%) Blood agar Lichenysin at 10°Ca fermentationb Production ± SD haemolysis synthetase gene A A30C01 - - + - + 25 ± 0.01 + + A30C11 - - + - + 21 ± 0.02 + + B30C11 - - + - + 33 ± 0.03 + + B30C12 - - + - + 42 ± 0.03 + + B30C13 - - + - + 52 ± 0.02 + + B30C14 - - + - + 45 ± 0.04 + + B55C02 - - + - + 42 ± 0.03 + + B55C11 - - + - + 53 ± 0.02 + + C30C01 - - + - + 57 ± 0.02 + + C30C11B - - + - + 48 ± 0.02 + + C55C01 - - + - + 57 ± 0.09 + + C55C02 - - + - + 66 ± 0.01 + + C55C11 - - + - + 39 ± 0.02 + + C55C12 - - + - + 74 ± 0.03 + + D30C03 - - + - + 75 ± 0.00 + + D30C14 - - + - + 37 ± 0.02 + + D55C01 - - + - - 71 ± 0.00 + + D55C02 - - + - + 56 ± 0.02 + + D55C03 - - + - + 62 ± 0.03 + + D55C04 - - + - + 53 ± 0.02 + + D55C11 - - + - + 57 ± 0.02 + + E30C01 - - + - + 72 ± 0.01 + + E30C04 - - + - + 63 ± 0.02 + + E30C11 - - + - + 51 ± 0.03 + + E55C01 - - + - + 71 ± 0.02 + + E55C02 - - + - + 42 ± 0.02 + + F30C01 - - + - - 69 ± 0.01 + + F30C02 - - + - + 63 ± 0.02 + + F30C11 - - + - - 68 ± 0.01 + + F30C14 - - + - + 53 ± 0.01 + + F55C01 - - + - + 34 ± 0.03 + + F55C02 - - + - + 54 ± 0.01 + + F55C12 - - + - + 62 ± 0.01 + + a Results from the microtitre plate reading at 570 nm and MPCA plate counting. b Results indicates no colour changes and gas production. c Results from tributyrin and spirit blue agar.

99

5.3.2 Pellicle formation and Congo red binding assay

The pellicle formation (air-liquid interface) and EPS production of B. licheniformis assayed by Congo red binding are shown in Figures 5.1 and 5.2. All of B. licheniformis isolates produced pellicles but the degree of pellicle formation for each strain are

varied based on the Congo red staining results at OD 490 nm.

A

A30C01 B30C11 C30C01 D55C11 E30C04 F30C11

B

Figure 5.1 Image (A) showing pellicles (air-liquid interface) formation by six B. licheniformis (A30C01, B30C11, C30C01, D55C11, E30C04 and F30C11) in TSB after 48 h of incubation at 37°C (The remaining 27 B. licheniformis isolates pellicles formation was not shown here). Image (B) showed pellicle production by B. subtilis as control.

100

100

90

80

70

60

50 EPS (%) 40

30

20

10

0

B. licheniformis isolates

Figure 5.2 The percentage of Congo red bound indicates the amount of EPS produced by B. licheniformis isolates. The absorbance was measured at

OD 490 nm. Mean and standard deviation were calculated from 4 replicates.

101

5.3.3 Heat resistance study at 72°C, 75°C and 80°C

The heat resistant study of selected B. licheniformis vegetative cells at 72°C, 75°C and 80°C are shown in Figure 5.3 to Figure 5.5.

9

8 B30C11 C30C11B 7 B30C12

log CFU/ml B55C02

6 15 30 45 60 75 time (s) Figure 5.3 The heat resistance of B. licheniformis vegetative cells at 72°C.

9

8 A30C01 B30C14 C55C01 7 log CFU/ml D55C04 E30C11

6 15 30 45 60 75 time(s)

Figure 5.4 The heat resistance of B. licheniformis vegetative cells at 75°C.

9

8 A30C01 B30C13 7 C55C11

log CFU/ml D55C01

6 15 30 45 60 75 time (s)

Figure 5.5 The heat resistance of B. licheniformis vegetative cells at 80°C.

102

5.4 DISCUSSION

B. licheniformis isolates from WPC80 powder are unable to grow at 10°C. Although they are thermotolerant and therefore able to survive pasteurisation, the inability to grow at 10°C means that they are unable to grow in the UF modules (inlet & outlet temperature is 10°C) of a whey manufacturing plant. This result agrees with Bergey’s manual (Logan & Vos, 2015) which stated the minimum growth for B. licheniformis is 15oC - 20°C depending on strains. UF membranes represent the largest surfaces area in a WPC manufacturing plant providing opportunity for extensive biofilm growth. However, most WPC manufacturing plants operate their UF membranes at 10°C. The inability of these isolates to grow at this temperature is important as this eliminates this section of a WPC plant as a potential zone for B. licheniformis growth.

Lactose is a sugar mainly found in milk. β-galactosidase is used to break down lactose to galactose and glucose that can be used as an energy source by many microbial cells. However, 28/33 of the B. licheniformis isolates from WPC80 powders were able to form biofilm in 1% RWPC80 without lactose or minerals on SS (Figure 4.5). It was postulated that these bacteria are unable to ferment lactose in the whey medium. Lactose fermentation analysis was done on all 33 B. licheniformis isolates confirming that 33 cannot ferment lactose. This finding is supported by Sadiq et al, (2016b). This observation seems surprising with bacteria that have originated from a lactose rich environment and suggests that some other organic material, such as protein, may provide the carbon source for these bacteria in a dairy environment. This new information is crucial in order to understand the survivability of B. licheniformis in the dairy industry without using lactose.

All 33 B. licheniformis produced protease indicted by clear zones along the streaked line on calcium caseinate agar. This degradation of protein suggests that these bacteria may be able to use protein as an alternative carbon source in the absence of a fermentable carbohydrate. This would enable them to survive and grow in a whey protein solution and as biofilm in WPC manufacturing plants. The ability of B. licheniformis to adapt to a carbohydrate depleted environment was demonstrated by Wiegand et al, (2013) where they demonstrated the importance of protease in fermentation.

103

The production of protease enzymes that may influence the quality of milk and milk products, is reported by Teh et al, (2012). The effect of the enzymes produced by these B. licheniformis on the sensory quality of Mozzarella WPC80 is unknown. WPC80 is a major food ingredient in secondary processing such as in UHT beverages, yoghurt and infant formula products that also include casein. The WPC80 concentrate level varies in these different products but the ability to produce protease is likely to influence the quality of all these products.

Another factor that potentially influences product quality and acceptability is the ability to convert nitrate and nitrite. Nitrate and nitrite are permitted at low levels (50 mg/kg calculated as nitrate ion) in food products (Food Standards Australia and New Zealand). Elevated levels of nitrite in the WPC manufacturing might be experienced by dairy manufacture and are a cause for concern by some dairy product customers (personal communication). In this study, 30 B. licheniformis isolates were able to convert nitrate into nitrite. Three isolates (D55C01, F30C01 and F30C11) were unable to reduce nitrate. The ability for these bacteria to convert nitrate to nitrite depends on availability of a substrate.

One of the unique features of B. licheniformis is the ability to produce lichenysin. Lichenysin is a lipopeptide that has antimicrobial activity. Madslien et al, (2013) identified lichenysin was produced by most B. licheniformis strains and has the cytotoxicity and biosurfactant effects. The production of lichenysin is indicted by haemolysis. The 33 B. licheniformis isolates from Mozzarella WPC80 samples produced haemolysis on Columbia sheep blood agar. The gene responsible for lichenysin production, Lichenysin synthetase A was detected in all 33 B. licheniformis isolates. The activity of what is now known to be lichenysin has been reported in a number of publications. Sayem et al, (2011) did an investigation on a broad spectrum anti biofilm exopolysaccharide from a marine strain of B. licheniformis that reduced the microbial colonisation of sponges. Batrakov et al, (2003) isolated B. licheniformis strains from drilling fluid and subsurface thermal water (56°C) that produce a lipopeptide that acted as an inhibitor for Corynebacterium variabilis adhesion on glass.

Two main biofilm components are microorganisms and EPS. Pellicle formation is a combination of bacteria and EPS forming a layer at a liquid-air 104

interface. Cells in pellicles are held together by an extracellular matrix consisting of exopolysaccharide, an amyloid-like fibres largely composed of protein (Kolodkin-Gal et al, 2012). Pellicle formation was observed for all 33 isolates of B. licheniformis when grown overnight in a microtitre plate with TSB as medium at 30°C, 37°C and 55°C (Chapter 3.3.1). This confirms their potential to form biofilm, which is believed to be the main source of contamination of WPC during manufacture.

Another indicator of EPS production and the potential to form biofilms is a Congo red binding assay. This involves treating a culture with Congo red which binds to the EPS produced by the bacteria. Ghafoor et al, (2011) used Congo red binding assay to investigate the glucose rich, Pel gene that associated with pellicles production of Pseudomonas aeruginosa (P. aeruginosa) between wild type and mutant. The mutant, Pel-negative did not produce pellicles and this was reflected in reduced Congo red binding compared with the wild type. The glucose rich gene, Pel is required for adherence and architecture at the later stages of biofilm maturity. Pel production indicates cell to cell interaction. This method was also used by Spiers et al, (2003) and Hay et al, (2009) for investigating the effects of, Pel production in artificial laboratory media (Luria bertani).

The results from the Congo red assay using the 33 B. licheniformis isolates showed that the amount of EPS production varied for each strain. There was no correlation between the amount of EPS produced during the Congo red assay and biofilm formation. It can be concluded that polysaccharide production is not the only factor responsible for biofilm formation of B. licheniformis isolated from WPC80. There is a high chance of B. licheniformis strains that produce more EPS might be more difficult to control by cleaning and sanitising in the dairy industry. B. subtilis was used as a control in this EPS assay as B. subtlis is the model in the Bacillus family for studying biofilm formation. Pellicle formation by B. subtilis when incubated in TSB showed a different wrinkled pattern from B. licheniformis as depicted in Figure 5.1. Roux et al, (2015) noted that poly-N-acetyl glucosamine is a major polysaccharide component of the B. subtilis biofilm matrix. This major carbohydrate component is essential for B. subtilis biofilms but may not be as critical for biofilms of B. licheniformis. This could be one explanation for the difference in pellical formation.

105

Several B. licheniformis isolates were subjected to heat treatment mimicking the thermalisation/pasteurisation temperature 72oC – 80°C to determine the heat resistance of their vegetative cells. Typically, B. licheniformis develops spores as a survival mechanism in nutrient depleted environments and these are very heat resistant to high temperature. The initial inoculum (8.0 ± 0.5 log CFU/mL) of tested B. licheniformis vegetative cells did not decrease even by 1 log reduction at the three temperatures tested; 72°C, 75°C and 80°C during 15 s, 30 s and 60 s heating intervals (Figure 5.3 – Figure 5.5). These results are in agreement with Yuan et al, (2012) in which they reported that the vegetative cells of 4 out of 5 B. licheniformis milk powder isolates cells were able to survive pasteurisation (72°C for 15s).

The B. licheniformis vegetative cells were heat resistant at these temperatures so D values could not be calculated. In addition, the B. licheniformis survival slightly increased after 30 s of heat treatment at both 72°C and 75°C. On the other hand, the 80°C temperature resulted in a slight decrease in B. licheniformis cell counts after 30 s up to 60 s (Figure 5.5). However, this temperature is out of the temperature range for the thermaliser (36oC – 77°C) used in the WPC80 manufacturing plant. Therefore, B. licheniformis vegetative cells are capable to survive this heat treatment providing opportunity to colonise the manufacturing plant and contaminate product.

106

5.5 CONCLUSIONS

B. licheniformis isolated from Mozzarella WPC80 powder is unable to ferment lactose but is able to grow in whey, a lactose rich medium. The fermentable carbon for the growth of B. licheniformis must come from either synergistic growth with lactose fermenting bacteria such as Lactobacillus species, also found in whey, or the use of protein as a source of carbon. Protease production by the B. licheniformis isolates suggests this use of protein as a substrate for growth is likely. Protease production will also influence product quality, in this instance, whey quality. Nitrate reduction by B. licheniformis is also likely to influence product quality. The B. licheniformis isolates from WPC grow as biofilms although the amount of EPS and biosurfactants that may influence biofilm formation does vary. The heat resistance of B. licheniformis vegetative cells isolated form WPC80 provides a new insight into the potential of these bacteria to survive conventional pasteurisation and contaminate the product.

107

______CHAPTER 6

______

BIOFILM AND SPORE FOMATION BY B. licheniformis IN SINGLE AND CO- CULTURE WITH L. plantarum IN 1% RWPC80 WITH LACTOSE AND MINERALS

108

6.1 INTRODUCTION

Biofilm in most natural environments consist of multi species rather than a single species of microorganism (Flint, 1997; Costerton, 1995; Anand et al, 2014). One of the factors that may enhance biofilm formation is the synergistic activities in co- cultures. A study by Zhao et al, (2013) demonstrated the relationship of co-culture growth between Geobacillus thermoglucosidans and A. flavithermus whereby Geobacillus thermoglusidans is unable to grow by itself. Abee et al, (2011) reported mixed species of L. monocytogenes and L. plantarum biofilms appeared to be more resistant to disinfectant treatments than single species biofilm or planktonic cells. Tang et al, (2009) showed an increase in biofilm formation where there was a combination of Klebsiella oxytoca isolates from dairy membrane on three types of UF membrane.

The 33 B. licheniformis isolated from Mozzarella WPC80 cannot ferment lactose (Chapter 5) and correlated to their API 50CHB kit test results negative for lactose utilisation (Table 3.8) so it is hypothesised that their growth must be supported by protein breakdown or synergistic activity with other microflora such as Lactobacillus species that can break down lactose to galactose and glucose for B. licheniformis use. This could be an important factor the growth of some strains of B. licheniformis in a whey product manufacturing plant.

One L. plantarum isolate (B30C21) was selected among three L. plantarum isolates from Mozzarella WPC80 (Table 3.9) for further investigation into a possible synergistic effect on selected B. licheniformis strains (E30C11 and F30C02), forming biofilm on SS. This B30C21 strain tested positive for lactose fermentation using 1% lactose broth (containing purple dye and reconstituted lactose) and API 50CHL kit.

109

6.2 MATERIALS AND METHODS

6.2.1 Source of isolates

For this chapter and onwards, two B. licheniformis and one L. plantarum (B30C21) were selected to investigate their synergistic study in forming biofilms on SS using 1% RWPC80 medium at 37oC. The properties of two selected B. licheniformis are described in Table 6.1

Table 6.1 The summary of the two selected B. licheniformis properties used in the synergistic study.

Biofilm on SS coupon Enzyme Lactose Lichenysin production breakdown 1% Isolate TSB 1% RWPC80 Protease Lipase 1% lactose Lch β- RWPC80 +lactose broth AA haemolysis + gene minerals E30C11 + + + + - - + + F30C02 - - + + - - + +

6.2.2 B. licheniformis E30C11 and F30C02 biofilm formation at 1, 2, 4, 8, 12 and 24 h in 1% RWPC80 with lactose and minerals at 37oC

In order to determine B. licheniformis biofilm formation within 24 h at their optimum growth temperature (37oC), the method used in Chapter 4.2.4 was used with the addition of shaking at 100 rpm to provide some kinetic energy as would be found in a flowing system. Each strain was tested separately in this trial. The SS coupons were retrieved at 1, 2, 4, 8, 12 and 24 h to quantify the amount of biofilm using a bead beating recovery (Chapter 4.2.4), serial 10-fold dilution in 0.1% peptone and growth on MPCA in duplicate at 37°C for 24 - 48 h. The individual colonies forming between 30 to 300 were counted for CFU per cm-2 quantification.

110

6.2.3 Co-culture study between B. licheniformis (E30C11 and F30C02) with L. plantarum (B30C21) biofilm growth 4, 8, 12 and 24 h in 1% RWPC80 with lactose and minerals

The same culture conditions, sampling and recovery of cells were used as for Section 6.2.2 but including 6 log CFU/mL L. plantarum strain (B30C21) with each selected 6 log CFU/mL B. licheniformis strain (E30C11 and F30C02). MPCA and de Man, Rogosa and Sharpe agar (MRSA) (Merck, BDH, Palmerston North, New Zealand) were used for plate counting B. licheniformis and L. plantarum respectively. Duplicate plates were prepared for each. MPCA plates were incubated at 37°C for 24 to 48 h. MRSA plates which are selective media for L. plantarum were incubated anaerobically in an anaerobic jar filled with gas pack (BBL™ GasPak™ Family, New Zealand) for 24 - 48 h at 30°C. The individual colonies on plates containing 30 to 300 were counted for CFU per cm-2 quantification.

6.2.4 Spore formation by B. licheniformis within biofilm in single and mixed populations

It is important to know the role of biofilm in the formation of spores of B. licheniformis as spores survive in the final WPC80 product. In addition to the enumeration for vegetative cells in Section 6.2.3, following the bead beating recovery (Chapter 4.2.4) the remaining samples after plating for vegetative bacteria were heated in a water bath at 80°C for 12 min prior to pour plating and incubation at 37°C. This heat treatment will eliminate the B. licheniformis vegetative cells and germinate the endospores (if any). Dhakal, (2013) noted that although their B. licheniformis were thermophilic, their growth and spore production was better at 37oC than 55oC, hence this was the incubation temperature used for this trial.

111

6.3 RESULTS

6.3.1 B. licheniformis E30C11 and F30C02 biofilm growth at 1, 2, 4, 8, 12 and 24h in 1% RWPC80 with lactose and minerals at 37oC

6 E30C11

5 F30C02

4 -2

3 Log CFUcm Log 2

1

0 0 5 10 15 20 25 30 Time (h)

Figure 6.1 Biofilm formation by E30C11 and F30C02 plotted against time. The mean and standard deviation were from triplicate (N=3) independent experiments.

The biofilm growth by two selected B. licheniformis strains (E30C11 and F30C02) is shown in Figure 6.1. There was a steady increase starting 2 h until 8 h for both strains indicating a generation time in the biofilm of approximately 1.5 h. E30C11 continued to increase cell numbers from 8 h until 12 h but F30C02 started to decline until 24 h of incubation.

112

6.3.2 Biofilm of a co-culture of (B. licheniformis and L. plantarum) at 4, 8, 12 and 24 h in 1% RWPC80 with lactose and minerals at 37oC

6 E30C11 F30C02 5 E30C11 + B30C21 F30C02 + B30C21 4 -2

3

Log CFULog cm 2

1

0 0 5 10 15 20 25 30 Time (h)

Figure 6.2 Counts of B. licheniformis E30C11 and F30C02 retrieved on MPCA from single species biofilm and mixed population biofilm with L. plantarum B30C21 in 1% RWPC80 with lactose and minerals. The mean and standard deviation were from triplicate (N=3) independent experiments.

5

B30C21 4 B30C21 + E30C11

-2 3 B30C21 + F30C02

2 Log CFU cm

1

0 0 5 10 15 20 25 30 Time (h)

Figure 6.3 Counts of L. plantarum B30C21 retrieved on MRSA from single species biofilm and mixed population biofilm with B. licheniformis E30C11 and F30C02 in 1% RWPC80 with lactose and minerals. The mean and standard deviation were from triplicate (N=3) independent experiments. 113

Biofilm formation of L. plantarum was measured by the recovery of bacteria on selective MRSA, incubated anaerobically from samples taken at 4, 8, 12 and 48 h at 30°C. This method allowed the selective isolation of L. plantarum (B30C21) cells from a mixed population.

The first four hours showed the dominance of a single population of B30C21 and B30C21 with F30C02 on SS. The highest bacterial CFU count on MRS agar -2 incubated anaerobically was 3.5 log10 CFU cm formed from the co-culture between B30C21 and F30C02. This observation supports the hypothesis that L. plantarum can grow well in a medium containing lactose as the main carbon source and therefore may produce surplus glucose and galactose for use by other microorganisms in a co- culture environment. Meanwhile, B. licheniformis cells retrieved from MPCA at 4 h -2 were about 2.5 log10 CFU cm either in single or co-culture environments, showing no benefit to these bacteria in a co-culture.

At 8 h, there were sharp declines in biofilm formation by the co-cultures of

E30C11 and F30C02 with B30C21. The decrease of approximately 1.5 log10 CFU cm-2 on MRSA may reflect the decrease in lactose concentration influencing the viability of the L. plantarum cells. A similar observation was noted on MPCA agar where there was a decrease of approximately 0.5 log CFU cm-2 possibly because of the production of lactic acid by the L. plantarum growth resulting in a reduction in the viability of B. licheniformis within the biofilm.

The biofilm formed from the co-culture of B. licheniformis and L. plantarum, -2 measured on MRSA, reached stationary phase after 8 h at 2.5 log10 CFU cm and remained at this level up to 24 h of incubation. Meanwhile, single and co-culture growth of B. licheniformis and L. plantarum measured on MPCA increased after 8 h -2 reaching the highest biofilm formation at 12 h (approximately 5 log10 CFU cm ) for

F30C02 and B30C21 as compared with E30C11 and B30C21 (an about 3.2 log10 CFU cm-2). However, biofilm formation of the co-culture of F30C02 and B30C21 started to decrease 1-fold from 12 h until 24 h. In contrast, biofilm formation by co-culture of

E30C11 and B30C21 still increased from 12 h until 24 h (approximately 5 log10 CFU cm-2).

114

6.3.3 Spore formation within co-culture population of biofilm

No spore formation by B. licheniformis was detected from single, mixed cultures (grown together with L. plantarum) in the biofilm at every time interval tested. Dhakal, (2013) reported the incubation time for spore production is 10 - 11 days and its production by B. licheniformis was greater at 37°C than at 55°C. The incubation time for the present trial was therefore much less than that reported by Dhakal, (2013) for spore formation.

115

6.4 DISCUSSION

B. licheniformis is unable to ferment lactose (Chapter 5.3.1) and therefore will need to use other sources, including protein such as in whey (α-lactalbumin and β- lactoglobulin) breakdown to amino acids as source of energy (Chapter 5.3.2). In this co-culture study for B. licheniformis (E30C11 and F30C02) and L. plantarum (B30C21) it was postulated that this would enhance biofilm formation, by the action of L. plantarum breakdown of lactose in the whey to glucose and galactose and that B. licheniformis would use the glucose as a carbohydrate supply. Although there was a small increased biofilm formation of B. licheniformis in 1% RWPC80 with lactose and minerals at 12 h, this was insufficient to indicate enhanced biofilm formation in this co-culture environment.

A co-culture of B. licheniformis (E30C11 and F30C02) and L. plantarum B30C21 did not increase biofilm formation of B. licheniformis in 1% RWPC80 with lactose and minerals medium at 24 h. B. licheniformis will form biofilms in the presence of whey and minerals and this is the most likely source of product contamination.

Stanley et al, (2003) reported that high glucose concentrations inhibit biofilm formation by B. subtilis and therefore growth is better in the planktonic form. Although their study used B. subtilis, it is possible that we are seeing a similar effect of glucose inhibition with B. licheniformis in this present trial with a decline in B. licheniformis after 8 h in the co-culture (Figure 6.2). This observation followed the L. plantarum single biofilm formation which declined from 4 h to 8 h indicating lactose depletion and therefore possible high glucose in the medium.

On the other hand, Assaf et al, (2015) reported that lactose triggers biofilm formation by Streptococcus mutans and Duanis-Assaf et al, (2016) concluded that the LuxS based quorum sensing governs lactose induced biofilm formation by B. subtilis.

The contamination of B. licheniformis in milk and milk products is mainly due to its ability to form spores which allows them to survive heat processes (Dhakal, 2013; Stoeckel et al, 2016). Spores from different Bacillus species vary in their attachment to solid surfaces. Husmark, (1993) discovered that attachment was higher in the spore state than the vegetative state in five different Bacillus species and B. cereus is the

116

most hydrophobic species. The attachment process is quite rapid and within 1 h, and a maximum level of adhesion is reached for specific spore concentrations. However, according to the results from Chapter 5, most B. licheniformis vegetative cells are heat resistant up to 80°C for 60 s and unable to reach 1 log reduction thus reflecting their potential to survive the heat treatment.

In general, the attachment of microorganisms to surfaces is a very complex process with many variables affecting the outcome. Although spores when they are present, dominate in the attachment process, Faille et al, (2014) found no correlation between biofilm and sporulation within biofilms of Bacillus species. Spore surfaces are hydrophobic due to proteins on their outer surface (Wiencek et al, 1990). The high relative hydrophobicity of Bacillus spores is reflected in a high degree of attachment and the hydrophobic interactions increase proportionally with temperature.

Seale et al, (2015) noted that “spores of B. licheniformis can survive pasteurisation but don’t appear to germinate and in processing milk lines therefore, they are not viewed as so much of a concern as spores from Geobacillus spp or A. flavithermus”. Although this finding was noticed in a milk processing plant, a similar pattern of B. licheniformis sporulation was observed in the whey environment in this study, where there was no evidence of spores in a biofilm state within 24 h. The spores might not be produced within 24 h although if biofilm and foulant were not completely removed during CIP and continue over successive runs, this could result in spore formation. The low spore counts seen in this trial for 6 WPC80 samples where the spore counts were below 100 CFU/g, supports this hypothesis that the sporulation of B. licheniformis does not occur readily. The presence of B. licheniformis in the final product may be due to the resistance of the vegetative cells to heat treatment and dehydration (Chapter 5.3.3). It would be interesting to see if these dairy isolates vary in their heat resistance compared to isolates from other sources. To control these bacteria in a dairy environment, alternative strategies to the standard heat treatment need to be considered.

117

6.5 CONCLUSIONS

B. licheniformis form biofilm on SS surfaces in a whey environment in single or mixed populations. A co-culture of B. licheniformis with L. plantarum does not increase the biofilm formation. Spore production by B. licheniformis within the biofilm at 4, 8, 12 and 24 h of incubation on SS at 37°C in whey does not occur over this time period. The heat resistance of the vegetative cells of B. licheniformis emphasises the importance of these bacteria in dairy manufacture through their ability to survive heat treatments such thermalisation/pasteurisation.

Locating the zone in the manufacturing plant where these biofilms grow will be important for their control. We know that B. licheniformis cannot grow at 10°C, the temperature used in the UF plant, therefore this cannot be a site of biofilm development for these bacteria. The most likely source of whey product contamination during whey processing is before UF process during clarification or in the cool zone of the thermalisation heat exchanger and separator where temperatures are approximately 37°C and the concentration of protein (0.6%) will support more biofilm growth than more concentrated solutions.

118

CHAPTER 7

THE POTENTIAL SOURCE OF B. licheniformis CONTAMINATION DURING MOZZARELLA WPC80 MANUFACTURE

119

7.1 INTRODUCTION

The steps involved in Mozzarella WPC80 samples manufacture are outlined in Figure 2.4. WPC80 manufacture includes chilling, clarifying, thermalisation (covering a wide temperature range) (Appendix 1), separation, UF/Diafiltration, evaporation, concentration and drying. The UF membrane plant provides a huge surface area as a potential for biofilm formation and bacterial growth. Tang et al, (2009) found Klebsiella strains isolated from cold UF membranes and some were shown to be good biofilm formers. However, they should be able to be controlled with normal cleaning procedures.

In Chapter 3, we reported that Mozzarella WPC80 powder samples contained high bacterial counts and showed the predominant microflora was the mesophilic thermotolerant bacterium B. licheniformis. B. licheniformis is unable to grow at the low temperatures (10°C) used in the UF membrane plants to concentrate the whey for WPC80 manufacture. Growth of B. licheniformis, if it occurs in the manufacturing plant, must occur earlier than the UF stage as it is these early stages where conditions (nutrient and temperature) are most suitable for the growth of B. licheniformis. Growth is most likely to occur as a biofilm.

Biofilm can cause blockages, reduce heat transfer and passing dairy liquid, such as milk, is contaminated with cells and enzymes released from the biofilm into the passing milk. This chapter aimed to determine the potential source of B. licheniformis contamination in Mozzarella WPC80 manufacture and the ability to form biofilm and cause protein damage and product contamination.

120

7.2 MATERIALS AND METHODS

7.2.1 Source of strains

Thirty-three B. licheniformis strains isolated from whey protein concentrate 80 (WPC80) samples as reported by Zain et al, (2016) were used in this study. The isolates were maintained on nutrient agar slants at 4oC and in glycerol beads stock at -80oC.

7.2.2 Liquid whey samples from pre-UF sites

Five liquid whey samples from different points in the whey processing plant, pre-UF (raw whey balance tank, clarifier, thermaliser and separator) (Table 7.1) were provided by the same whey processing plant reported in Chapter 3.2.1.

Table 7.1 Liquid whey samples from pre-UF process of WPC80 powder.

Pre-UF liquid whey samples 1. Raw whey balance tank (A) 2. Clarifier (B) 3. Balance tank/thermaliser (C) 4. Post module/thermaliser (D) 5. Separator (E)

7.2.3 Isolation, identification, characterisation and biofilm formation of B. licheniformis isolates from pre-UF liquid whey samples

One mL of each liquid whey samples was diluted in 90 mL of sterile 0.1% peptone water (pH 7.0) and mixed using a vortex mixer. Samples were tested for microbial content using milk plate count agar (MPCA) (Merck, BDH, Palmerston North, New Zealand) and incubated at 37°C. The pure cultures grown on MPCA as single isolated colonies were screened using basic colony morphology and Gram staining as presumptive B. licheniformis (lichen-like colonies). Identification was confirmed by biochemical testing and partial 16S rDNA gene sequencing (Flint et al, 1999b).

121

Characterisation of presumptive B. licheniformis from pre-UF liquid whey samples were done on lactose fermentation, protease, lipase, nitrate, lichenysin and pellicles production as described in Chapter 5.2.

The confirmed B. licheniformis isolates were tested for biofilm formation on microtitre plates using TSB (Oh et al, 2007) (Figure 4.2.2) and subsequently on SS as described in section 4.2.4.

7.3 RESULTS

7.3.1 Isolation, identification and characterisation of B. licheniformis isolates from pre-UF liquid whey samples

The optimum temperature for B. licheniformis was 37oC based on microtitre plate results (Chapter 4) and our findings showed that they formed biofilm best with 1% RWPC80 with lactose and minerals (Chapter 4). This information indicates that the source of B. licheniformis contamination in whey manufacture is most likely before the UF process where the whey protein before concentration is about 0.6% and lactose concentration 4.5%.

To test our hypothesis, and identify the most likely zones in the manufacturing plant for biofilm growth of B. licheniformis, five liquid whey samples from different sites; raw whey balance tank, clarifier, thermaliser and separator were tested for B. licheniformis. Seven B. licheniformis isolates were found (2 isolates from the raw whey balance tank; A2 and A3), (1 isolate from clarifier; B2), (1 isolate from the thermaliser; D2) and 3 isolates from separator E1, E3 and E4) and identified by Gram- staining, colony and pellicle formation, API 50CHB testing and 16S rDNA gene sequencing as described in Chapter 3.

Further tests on their characteristics were done as described in Chapter 5.2. Figure 7.1 shows their colony formation on MPCA (A), β-haemolysis on Columbia blood agar after 48 h (B) (as indication of lichenysin production) and pellicle formation (C). Table 7.3 tabulated the summary of their characteristics as compared to E30C11 and F30C02 derived from Mozzarella WPC80 powder samples.

122

A B

C

A2 A3 B2 D2 E1 E3 E4

Figure 7.1 Phenotypic characteristic of pre-UF B. licheniformis A) colony formation on MPCA, B) β-haemolysis on Columbia with sheep blood agar, C) pellicle formation of A2, A3, B2, D2, E1, E3, and E4 strains after 2 days of incubation at 37°C in TSB.

A summary of the characteristics of the B. licheniformis isolates recovered from the pre-UF part of the manufacturing plant are listed in Table 7.2 and compared with E30C11 and F30C02 isolates from WPC80 powder samples.

123

Table 7.2 Summary of B. licheniformis characteristic from pre-UF sites compared to E30C11 and F30C02 from the WPC80 powder samples.

Characteristic E30C11 F30C02 A2 A3 B2 D2 E1 E3 E4 Colony morphology Dry Mucoid Mucoid Mucoid Mucoid Mucoid Mucoid Mucoid Mucoid Opaque Opaque Opaque Opaque Opaque Opaque Opaque Opaque Opaque Irregular Irregular Irregular Irregular Irregular Irregular Irregular Irregular Irregular edges edges edges edges edges edges edges edges edges Cell morphology Rod Rod Rod Rod Rod Rod Rod Rod Rod Gram staining Gram Gram Gram Gram Gram Gram Gram Gram Gram positive positive positive positive positive positive positive positive positive Growth temperature Growth at 10oC ------Growth at 30oC + + + + + + + + + Growth at 37oC + + + + + + + + + Growth at 55oC + + + + + + + + +

Lactose fermentation ------Pellicles formation + + + + + + + + + Enzyme production Protease + + + + + + + + + Lipase ------Lipopeptide: Lichenysin + + + + + + + + + API 50CHB 99% 99% 99% 99% 99% 99% 99% 99% 99%

124

7.3.2 Biofilm formation by pre-UF B. licheniformis isolates on plastic and SS

A preliminary biofilm study using the microtitre plate assay in TSB showed A2 and D2 formed strong biofilm. The remaining five B. licheniformis isolates did not form biofilm and their absorbance in the microtitre plate assay were similar to the negative control (sterile TSB without inoculum) (Figure 7.2).

4

3

2

1 Absorbanceat 570 nm

0 A2 A3 B2 D2 E1 E3 E4 pre-UF B. licheniformis isolates

Figure 7.2 The microtitre plate assay results for biofilm formation in TSB by 7 B. licheniformis isolates taken before UF process. A2 and A3 isolates were derived from the raw whey balance tank (A), B2 isolates from clarifier (B), D2 from thermaliser and E1, E3 and E4 from separator. Only A2 and D2 showed strong biofilm formation on the polystyrene microtitre plate surface. The mean and standard deviation were from triplicate (N=3) independent experiments.

125

Further tests were done on the seven B. licheniformis isolates to examine their biofilm formation using three different media on SS and the results showed that they -2 formed good biofilm in all 3 different media between 2.16 – 4.78 log10 CFU cm (Figure 7. 3). The potential source of B. licheniformis contamination in the Mozzarella WPC80 powder is shown in the Figure 7.4.

TSB 5

1% RWPC 80 4

1% RWPC 80 lactose and 2 3 minerals

Log CFU cm-LogCFU 2

1

0 A2 A3 B2 D2 E1 E3 E4

pre-UF B. licheniformis isolates

Figure 7.3 Biofilm formation on SS in 3 different media by 7 B. licheniformis isolates. A2 and A3 isolates were derived from the raw whey balance tank (A), B2 isolates from clarifier (B), D2 isolate from thermaliser (C, D) and E1, E3 and E4 from separator (E). The mean and standard deviation were from triplicate (N=3) independent experiments.

126

Mozzarella Raw Clarifier Thermaliser Separator Chiller Ultrafiltration/ process whey Diafiltration balance tank 40°C 14°C 35°C 36-77°C 40-44°C 8°C 10°C

(A) (B) (C) (D) (E)

Dryer (air inlet) Concentrate Evaporator Evaporator handling preheat

>200°C 40°C 51-53°C 25°C

Figure 7.4 The diagram showed the potential source of contamination of B. licheniformis in the Mozzarella WPC80 powder; Raw whey balance tank (A), Clarifier (B), Thermaliser (C, D) and Separator (E). The highlighted letters with yellow colour indicates the most likely zone for biofilm growth and B. licheniformis contamination.

127

7.4 DISCUSSION

The objective of this chapter was to determine the possible source of predominant B. licheniformis contamination in a Mozzarella WPC80 manufacturing plant. Traditionally, microbial contaminants were believed to grow on the membrane surfaces of the UF plant as this represents the largest surface area in the plant as some plants operate their UF processes at 40°C (Lehmann, 1995). Changes from hot to cold UF have reduced the growth potential for bacteria on the membrane surfaces.

Seven B. licheniformis isolates were isolated from pre-UF liquid whey samples. This was confirmed by microscopy, colony and pellicle formation, biochemical and bio-molecular tests and compared with information from Chapters 3 and 5 (Zain et al, 2017). Their characteristics were in agreement with the B. licheniformis isolates from WPC80 powder (Logan & Vos, 2015). Their ability to attach and form biofilm was screened on microtitre plates and 2 (A2 and D2) isolates from 7 showed strong biofilm formation while the remaining 5 were non biofilm former (ODc value at 0.5). Further investigation using three different media; TSB, 1% RWPC80, 1% RWPC80 with lactose and minerals revealed that the 7 B. licheniformis were able to form biofilm on SS and the results showed that they formed good biofilm -2 in all 3 different media between 2.16 – 4.78 log10 CFU cm (Figure 7.3) (Aulikki, 2017). B. licheniformis biofilm formations were studied using the microtitre plate assay (CV staining) and cell enumeration on SS in this study. These results followed those in Chapter 4 for the WPC80 powder isolates that biofilm formed better on SS than the plastic surface of microtitre plate. The entire manufacturing plant containing SS surfaces provides potential sites for biofilm formation of B. licheniformis and this represents a much larger surface area than any other parts, such as rubber seals. However, contributing factors such as medium concentration, minerals content and temperature will affect the biofilm growth on SS.

In Chapter 4, it was shown that the predominant microflora B. licheniformis would not grow in the membrane plant because of the low temperature (10°C) and must be growing elsewhere. In addition, the effect of whey concentration on growth, water activity of concentrated product and the importance of cations further limited the opportunity for growth on the UF membrane.

128

Contamination of dairy products is mostly due to bacteria being released from biofilm in the processing plant rather from the farm itself. Crielly et al, (1994) identified B. licheniformis as the common species isolated in raw milk and contaminating the dairy production chains. B. licheniformis was also the predominant bacteria colonising UF plant operated at 40oC used for standardising milk used in cheese manufacture (Lehmann, 1995). The source of biofilms of thermotolerant bacteria such as B. licheniformis is believed to be before the UF zone where the whey protein concentration is about 0.6%. This bacterium is thermotolerant and capable of growing in sections of dairy manufacturing plant where temperatures reach 40oC– 65°C (Burgess et al, 2010). Furthermore, because it is a spore former it is difficult to eliminate using food processing techniques. Spore forming bacteria such as Bacillus species can survive and colonise dairy manufacturing plant as a biofilm, releasing enzymes and bacteria to contaminate milk products (Burgess et al, 2010; Yuan et al, 2012).

Miller et al, (2015) managed to isolate Bacillus species spores in sweet whey, non-fat dry milk, acid whey and WPC80 and the highest frequency of B. licheniformis spores were detected in all whey products. It is important to locate the zone in the whey manufacturing plant where this bacterium survives and forms biofilm in order to be able to develop strategies for the control of this bacterium in a whey processing plant.

However, very little information is available on the source of contamination of these ubiquitous bacteria in whey processing plant. A study by Sadiq et al (2016a); Cook & Sandeman (2000) and Dhakal, (2013) reported B. licheniformis was the dominant thermophilic spore former in dairy products where the latter concluded that B. licheniformis survives in manufacturing plants in various forms or genotypes. This heterogeneity may assist in the survival in dairy environments and dairy products. The safety and quality of dairy products relies on heat treatments such as pasteurisation.

129

7.5 CONCLUSIONS

Biofilm formation by B. licheniformis isolates from WPC80 varies between strains and is dependent on the substrate surface with SS supporting more growth than polystyrene. TSB, whey concentration of 1% RWPC80 with and without lactose and minerals supports biofilm formation of 7 B. licheniformis from pre-UF sites unlike the B. licheniformis isolates from WPC80 powder where some of the isolates grew a better biofilm in 1% RWPC80 with lactose and minerals.

The most likely source of whey product contamination during whey processing in this case study is likely to be before UF (Figure 7.4) where the concentration of protein is 0.6%. Likely areas include, during clarification, in the cool zone of the thermalisation heat exchanger where temperatures are approximately 37°C and the separator where the temperatures around 40oC - 44°C. This combination will support more biofilm growth than more concentrated solutions. This suggestion is supported with the results from isolation, identification, characterisation and biofilm formation of 7 B. licheniformis isolates from liquid pre-UF samples that have been presented and discussed in this chapter.

130

CHAPTER 8

FINAL DISCUSSION AND CONCLUSIONS

131

8.1 DISCUSSION

In this study, microorganisms that contaminated six Mozzarella WPC80 powders come from one WPC80 manufacturing plant. The six Mozzarella WPC80 powders used for this study contained 4-5 log10 CFU/g total aerobic mesophiles and < 100 CFU/g mesophilic aerobic spores. A study by Okuku et al, (2017) noted that aerobic mesophiles were the dominant contaminants in WPC80 powder after 18 months’ storage compared to WPC34 which contained higher yeast and mould counts. However, the different whey sources, manufacturing process, storage conditions are likely to have contributed to the different results in study.

In this study, bacteria were identified using biochemical and molecular techniques. The predominant contaminants were B. licheniformis. The culturable population only was considered in this study as these are responsible for the quality grading of WPC resulting in consumer rejection of product and a cost to the dairy manufacturing company. Dairy product contamination is believed to be enhanced by the growth of biofilms in the manufacturing plant resulting in high microbial numbers from bacteria being released from the biofilm into the product. Biofilms may develop on any surface where the conditions are suitable for bacterial growth (Costerton 1995; Donlan 2002; Flint et al, 2001). These conditions include the availability of nutrients and suitable temperature and physical conditions such as a suitable substrate and flow conditions to allow the attachment and growth of microorganisms. In the dairy industry, biofilms develop on SS surfaces from which they detach and contaminate product. For example, S. thermophilus biofilm on UF membranes and SS pipework of a standard whey processing plant resulted in whey powders containing bacteria at > 4 log CFU/g (Flint, 1998).

The bacteria contaminating dairy manufacturing plant and dairy product are generally assumed to originate from the raw milk. Gram positive bacteria dominate the population in fresh raw milk but Gram negative bacteria become dominant as the milk is stored under refrigeration before entering a dairy manufacturing plant. Dairy manufacture selects for the more heat resistant Gram positive bacteria and propagates these through the development of biofilms in the manufacturing plant (Raats et al, 2011). Each Gram positive genera will vary in the requirements for growth within a

132

manufacturing plant. The conditions leading to B. licheniformis biofilm growth in a modern WPC manufacturing plant have not been determined until this study.

Biofilm formation by B. licheniformis, the dominant microflora in the WPC80 product analysed in this trial, was initially evaluated at three different temperatures; 30°C, 37oC and 55°C using a microtitre plate assay. The degree of attachment and biofilm formation of B. licheniformis isolates was determined after 24 h incubation in TSB medium in static conditions. The results were divided into 3 categories; none, moderate and strong biofilm formation. B. licheniformis biofilm formation occurred preferentially at 37°C with 9/33 isolates formed strong biofilm. Meanwhile, 24/33 B. licheniformis isolates were identified as unable to form biofilm at the three different temperatures.

Dairy practices using 30°C and 55°C do encourage the growth of some Bacillus species however Zhao et al, (2013) reported that incubation at 65°C is favoured by thermophilic bacteria such as Geobacillus species and A. flavithermus. Dhakal, (2013) noted that their B. licheniformis strain isolated from milk was reluctant to attach microtitre plates using TSB medium. The microtitre plate assay is a convenient rapid method to screen microbial isolates for biofilm formation. However, the suitability of this test has been questioned for some bacterial species (Peeters et al. 2008). It appears that from the present study that the microtitre plate assay is not suitable for B. licheniformis. This agrees with Dhakal, (2013).

In the present trial, there were two factors to consider in the assessment of the ability of the B. licheniformis isolates to form biofilm. One was the microtitre plate surfaces which Dhakal, (2013) had reported were unsuitable for B. licheniformis biofilm testing and the second was the lab medium compared with whey. Therefore, the assessment of biofilm formation was done on SS surfaces in whey medium and this was compared with SS surfaces in TSB.

Reconstituted whey medium (Tang, 2011) was used at three different concentrations (1%, 5% and 20%) to mimic the composition at three different stages in WPC manufacture. In this study, most biofilm formed at the 1% concentration with lactose and minerals (removed later in the manufacturing process) added. This eliminates a large section of the manufacturing plant (UF and evaporation) as sites to

133

support biofilm formation of B. licheniformis. Some strains were unable to form biofilm in 1% WPC without minerals. Minerals are removed later in the WPC manufacturing process, therefore any B. licheniformis growth is most likely to occur before dialysis.

The inability B. licheniformis to grow at 10°C (the temperature of the UF process) as discussed in Chapter 5 is further evidence to eliminate UF as a section of plant where these bacteria may propagate. This is important as the surface area of an UF plant is large and potentially a site for microbial colonisation.

The most likely sites for B. licheniformis growth in a WPC manufacturing plant are believed to be the clarifier (to remove particles), thermaliser (to reduce microbial load in the cheese whey) and separator (to remove fat). All these unit operations have zones at suitable temperatures for the growth of B. licheniformis. This was supported with the isolation of 7 B. licheniformis isolates retrieved from these areas of the manufacturing plant. Three from 7 isolates were from the separator which operates at 45°C. CIP and other control measures such as altering temperatures, to reduce biofilm formation and reduce contamination of WPC with B. licheniformis need to focus on these areas.

Dhakal, (2013) concluded that an important consideration in the survival and dispersal of B. licheniformis in milk and milk products is the ability to produce spores which survive heat treatment and can germinate in suitable zones of a dairy manufacturing plant. Spore counts of B. licheniformis in raw milk are usually < 50 CFU per mL (Crielly et al, 1994). B. licheniformis was reported as the dominant thermophilic spore forming bacteria in dairy products (Cook & Sandemann, 2000; Dhakal, 2013). Ronimus et al, (1997) revealed interesting results on thermophilic spores of B. licheniformis and Geobacillus species in that they survived over 90 years in New Zealand milk powder.

In the present study, no spore production was detected in biofilm grown in 1% RWPC80 with minerals on SS at 4, 8, 12 and 24 h of incubation. This observation may be because the spore formation of B. licheniformis requires more than 24 h. As manufacturing plants are cleaned on a regular basis, well before 24 h, it is curious that spores can form in the manufacturing plant. The most likely explanation is that

134

residual bacteria and spores may persist in a mature biofilm after cleaning and continue to propagate in successive manufacturing runs, allowing them time to produce spores (Witthuhn et al, 2011). The B. licheniformis vegetative cells were heat resistant when tested at 72°C, 75°C and 80°C with < 1 log reduction from the initial (~8 ± 0.5 log CFU/mL) population (Chapter 5).

The characterisation of the isolates of B. licheniformis surprisingly showed that they did not ferment lactose. This was not expected for bacteria isolated from a dairy system, especially as they were shown to need lactose to support the best biofilm growth. The bacteria must obtain carbohydrate and one possibility is the use of protein in the whey (α-lactalbumin and β-lactoglobulin).

All isolates in this trial produced protease enzyme (not lipase), which is a concern in terms of spoilage of whey. Enzymes like protease and lipase do influence the quality of milk and milk products by altering the sensory qualities such as texture, taste and aroma. They will also affect the nutritional value (Chen et al, 2003; Teh et al, 2012; Teh et al, 2013). Several other Bacillus and Paenibacillus species produce a variety of proteases, lipases and phospholipases that affect the quality of dairy products (Lücking et al, 2013) although it is often the psychrotrophic population that is responsible for much of the enzyme deterioration of dairy products. A study by Raats et al, (2011) concluded that an estimated 10% loss of protein and fat in milk was because of enzymes produced by psychotropic bacteria. Psychrotrophs produce enzymes that cause defects in cheese by lowering the yields and producing off flavours, especially rancidity via lipase production (Champagne et al, 1994). Another factor that potentially influences product quality and acceptability is the ability B. licheniformis to convert nitrate and nitrite. Nitrate and nitrite are permitted at low levels (50 mg/kg calculated as nitrate ion) in food products (Food Safety Australia and New Zealand). Elevated levels of nitrite in the WPC manufacturing might be experienced by dairy manufacture and are a cause for concern by some dairy product customers.

Although not the main focus of this study, P. glucanolyticus was isolated from the WPC80 powder samples and is a potential spoilage bacterium. Paenibacillus species grow at psychrotrophic temperatures and are therefore able to grow during whey UF which operates at 10°C (Vithanage et al, 2014). A study by Huck et al, 135

(2007) identified Paenibacillus as a bacterium responsible for significant spoilage that limited HTST-pasteurised milk product shelf life. This was supported by a study by Ivy et al, (2012) on Paenibacillus as the predominant psychrotolerant spore former in fluid milk. Driehuis et al, (2016) showed Paenibacillus polymyxa spores to be present in dairy cow milk, faeces and silage.

The focus of the present study was on B. licheniformis as this was the predominant isolate in the WPC tested. As mentioned earlier, one of the surprising outcomes from this study was the inability of B. licheniformis to utilise lactose. One possible source of carbohydrate could be the glucose released from lactose utilisation by other bacteria in WPC. Synergistic interplay between B. licheniformis with L. plantarum was hypothesised to increase the magnitude of biofilm formation as the latter able to convert lactose to glucose. However, the biofilm in co-culture environment with these two species did not increase biofilm formation by either bacterium.

There are other aspects that were considered in understanding the formation of biofilm in a WPC manufacturing plant. These include the ability to produce EPS (an essential component of a biofilm) and the production of biosurfactants (which have been reported for B. licheniformis).

EPS production was detected in all 33 B. licheniformis isolates including 7 pre-UF B. licheniformis isolates. However, there was no correlation between the amount of EPS produced during the Congo red assay and biofilm formation. It can be concluded that polysaccharide production is not the only factor responsible for biofilm formation of B. licheniformis isolated from WPC80. B. licheniformis is reported to produce a biosurfactant, lichenysin. The 33 B. licheniformis including 7 pre-UF B. licheniformis isolates from this study were all shown to contain the gene for lichenysin production. The haemolysis of sheep blood, shown for the B. licheniformis isolates in this study, is an indicator of lichenysin production. However, attempts to isolate lichenysin were unsuccessful. This may be due to difficulties with the method used or the haemolysis results may not be reflecting lichenysin production. If that is true, we assume that under our experimental conditions, the gene for lichenysin production was not expressed and future work needs to look at conditions that may result in the expression of this gene. Further experiments on different extraction methods from B. 136

licheniformis within biofilm in whey environments need to be addressed. This is important as biosurfactants have been shown to help in the dispersal of B. subtilis from a mature biofilm (Branda et al, 2001) and therefore may be able to be used to control B. licheniformis.

This study differs from a processing facility in that batch experiments rather than a continuous process was used. Continuous experiments with a continuous supply of fresh nutrient would be more representative of a WPC processing plant. To confirm the most likely sources of contamination with B. licheniformis in WPC processing, an intensive swabbing/sampling at different zones within manufacturing plants (plate heat exchanger of thermaliser and clarifier/separator) would provide useful information to help in targeting sites for controlling contamination. For routine monitoring, testing samples such as separator sludge for testing at 37°C rather than the current currently only 30°C and 55°C, would have a better chance of isolating B. licheniformis.

The strategies to control B. licheniformis in WPC processing include more frequent cleaning of the thermaliser or running two thermaliser, each for short periods so the growth of B. licheniformis is limited. The flow rate (turbulent flow) could be controlled as many studies reports that increased flow enhances the release of bacteria from biofilms. Surface modification of stainless steel in terms of hydrophobicity and surface charges, silver impregnation and antimicrobial coatings (bacteriocin) might help reduce the attachment of microorganisms. Modifications to the CIP practices including a greater use of surfactants, enzymes and chelating agents may assist in controlling biofilms of B. licheniformis (Anand et al, 2014).

137

8.2 CONCLUSIONS

B. licheniformis has the potential to be a major contaminant of WPC. Biofilm growth is the most likely source of product contamination and this can only occur in some very specific parts of the manufacturing plant where conditions such as temperature and salt content are suitable. These parts of a WPC manufacturing process (clarifier, thermaliser and separator) need to be the focus for any modifications to the manufacturing process to control the biofilm growth of these bacteria. Protease production makes these bacteria a concern for spoilage of the WPC and any consumer products manufactured from this product. The conditions resulting in spore production, biosurfactant release and EPS production need further investigating.

8.3 HIGHLIGHTS OF THE STUDY

1. The best temperature for biofilm formation of B. licheniformis was at 37°C. 2. SS was the best surface for biofilm formation compared to plastic for B. licheniformis. 3. 1% reconstituted WPC80 with salts, supports biofilm formation by B. licheniformis. 4. B. licheniformis is unable to grow under the conditions present in a whey UF plant operating at 10°C. 5. Vegetative cells of B. licheniformis isolated from WPC80 survive pasteurisation temperatures. 6. The source of B. licheniformis contamination in WPC80 processing plant is postulated at thermaliser and separator regions. 7. B. licheniformis isolates were unable to ferment lactose but able to grow in a dairy environment.

138

8.4 FUTURE RECOMMENDATIONS

1. Synergistic studies with Paenibacillus species and thermophilic bacteria such as A. flavithermus to determine if Paenibacillus species can provide a fermentable carbohydrate from lactose for B. licheniformis growth. 2. Exploring the conditions for lichenysin production within biofilm and how this may influence biofilm growth. 3. Determining the conditions that result in spore formation in biofilms of B. licheniformis. 4. An in-depth study on CIP practise around the critical areas (before UF process) and its effects on B. licheniformis cell counts. 5. Develop ways to control biofilm formation and growth of spore-formers in WPC production.

139

REFERENCES

Abee, T., Kovacs, A. T., Kuipers, O. P., & Veen, S. V. D. (2011). Biofilm formation and dispersal in Gram-positive bacteria. Current Opinion in Biotechnology. 22, 172-179. Alvarez-Ordóñez, A., Begley, M., Clifford, T., Deasy, T., Considine, K., O’Connor, P., Ross, R. P., & Hill, C. (2014). Investigation of the antimicrobial activity of Bacillus licheniformis strains isolated from retail powdered infant milk formulae. Probiotics & Antimicrobial Proteins, 6, 32-40. Anand, S., Singh, D., Avadhanula, M., & Marka, S. (2014). Development and control of bacterial biofilms on dairy processing membranes. Comprehensive Reviews in Food Science and Food Safety, 13, 18-33. Anuradha, S. N. (2010). Structural and molecular characteristics of Lichenysin and its relationship with surface activity. Biosurfactants in Advances in Experimental Medicine and Biology, 672, 304-315 Aouadhi, C., Maaroufi, A., & Mejri, S. (2013). Incidence and characterisation of aerobic spore-forming bacteria originating from dairy milk in Tunisia. International Journal of Dairy Technology, 67 (1), 95-102. Assaf, D., Steinberg, D., & Shemesh, M. (2015). Lactose triggers biofilm formation by Streptococcus mutans. International Dairy Journal, 42, 51-57. Aulikki, K. (2016). Early stages biofilm formation of the facultative thermophilic sporeforming Bacillus licheniformis on stainless steel surface. http://lup.lub.lu.se/student-papers/record/8881137 Bansal, B., & Chen, X. D. (2006). A critical review of milk fouling in heat exchangers. Comprehensive Reviews in Food Science and Food Safety, 5, 27- 33. Batrakov, S. G., Radionova, T. A., Esipov, S. E., Polykov, N. B., Sheichenko, V. I., Shekhovtsova, N. V., Lukin, S. M., Panikov, N. S., & Nikolaev, Y. A. (2003). A novel lipopeptide, an inhibitor of bacterial adhesion from the thermophilic and halotolerant subsurface Bacillus licheniformis strain 603. Biochimica et Biophysica Acta (BBA)- Molecular & Cell Biology of Lipids, 1634(3), 107- 115. Bellona, C., & Drewes, J. E. (2005). The role of membrane surface charge and solute physicochemical properties in the rejection of organic acids by NF membranes. Journal of Membrane Science, 249, 227-234. Branda, S. S., González-Pastor, J. E., Ben-Yehuda, S., Losick, R., & Kolter, R. (2001). Fruiting body formation by Bacillus subtilis. PNAS, 98(20), 11621- 11626. Buehner, K. P., Anand, S., & Garcia, A. (2014). Prevalence of thermoduric bacteria and spores on 10 Midwest dairy farms. Journal of Dairy Science, 97(11), 1-8. Burgess, S. A., Brooks J. D., Rankonjac, J., Walker, K. M., & Flint, S. H. (2009). The formation of spores in biofilms of Anoxybacillus flavithermus. Journal of Applied Microbiology. 107(3), 1012-1018.

140

Burgess, S. A., Lindsay, D., & Flint, S. H. (2010). Thermophilic bacilli and their importance in dairy processing. International Journal of Food Microbiology, 144, 215-225. Burgess, S. A., Flint, S. H., & Lindsay, D. (2013). Characterization of thermophilic bacilli from a milk powder processing plant. Journal of Applied Microbiology, 116, 350-359. Busscher, H. J., Bellon-Frontaine, M. N., Mozes, N., & Van der Mei, H. C., (1990). Deposition of Leuconostoc mesentroides and Streptococcus thermophilus to solid substrata in a parallel plates flow cell. Biofouling, 2, 55-63. Busscher, H. J., Vanderkuijbooij, M., & Van der Mei, H. C. (1996). Biosurfactants from thermophilic dairy Streptococci and their potential role in the fouling control of heat exchanger plates. Journal of Industrial Microbiology, 42, 405- 412. Chamberland, J., Lessard, M. H., Doyen, A., Labrie, S., & Pouliot, Y. (2017a). A sequencing approach targeting the 16S rRNA gene unravels the biofilm composition of spiral-wound membranes used in the dairy industry. Dairy Science & Technology. 96(6), 827-843. Chamberland, J., Lessard, M. H., Doyen, A., Labrie, S., & Pouliot, Y. (2017b). Biofouling of ultrafiltration membrane by dairy fluids: Characterization of pioneer colonizer bacteria using a DNA metabarcoding approach. Journal of Dairy Science. 100 (2), 981–990. Champagne, C. P., Laing, R. R., Roy, D., Mafu, A. A., & Griffiths, M. W. (1994). Psychrotrophs in dairy products; their effects and their control. Critical Reviews in Food Science and Nutrition, 34, 1-30. Chen, L., Daniel, R. M., & Coolbear, T. (2003). Detection and impact of protease and lipase activities in milk and milk powders. International Dairy Journal. 13, 255-275. Cook, G. M., & Sandeman, R. M. (2000). Sources and characterisation of spore- forming bacteria in raw milk. Australian Journal of Dairy Technology, 55, 119-126. Coronel-León, J., Pinazo, A., Pérez, L., Espuny, M. J., Marqués, A. M., & Manresa, A. (2017). Lichenysin-geminal amino acid-based surfactants. Synergistic action of an unconventional antimicrobial mixture. Colloids and Surfaces B: Biointerfaces, 149, 38-47. Costerton, J. W. (1995). Overview of microbial biofilms. Journal of Industrial Microbiology, 15, 137-140. Crielly, E. M., Logan, N. A., & Anderton, A. (1994). Studies on the Bacillus flora of milk and milk products. Journal of Applied Bacteriology, 77, 256-263. Daane, L. L., Harjono, I., Barns, S. M., Launen, L. A., Palleroni, N. J., & Haggblom, M. M. (2002). PAH-degradation by Paenibacillus spp. and description of Paenibacillus naphthalenovorans sp. nov., a naphthalene-degrading bacterium from the rhizosphere of salt marsh plants. International Journal of Systematic and Evolutionary Microbiology, 52, 131-139.

141

De Jonghe, V., Coorevits, A., Block, J. D., Coillie, E. V., Grijspeerdt, K., Herman, L., De Vos, P., & Heyndrickx, M. (2010). Toxigenic and spoilage potential of aerobic spore-formers isolated from raw milk. International Journal of Food Microbiology, 136, 318-325. De Wit, J. N. (2001). Whey protein concentrates: manufacture, composition and applications. Industrial Proteins, 9(3), 3-5. Delaney, R. A. M. (1976). Composition, properties and uses of whey protein concentrate. Journal of the Society of Dairy Technology, 29, 91-101. Delorme, C. (2008). Safety assessment of dairy microorganisms: Streptococcus thermophilus. International Journal of Food Microbiology, 126(3), 274-277. Dhakal, R. (2013). Molecular characterization of B. licheniformis from dairy products. The University of Queensland, Australia, Thesis. Dhakal, R., Chauhan, K., Seale, B. R., Deeth, H. C., Pillidge, C. J., Powell, I. B., Craven, H., & Turner, M. S. (2013). Genotyping of dairy Bacillus licheniformis isolates by high resolution melt analysis of multiple variable number tandem repeat loci. Food Microbiology, 34, 344-351. Dogan, Z., Weidendorfer, K., Müller-Merbach, M., Lembke, F., & Hinrichs, J. (2009). Inactivation kinetics of Bacillus spores in batch and continuous heating systems. LWT- Food Science and Technology, 42, 81-86. Donlan, R. M. (2002). Biofilms: Microbial life on surfaces. Emerging infectious diseases, 8(9), 881-890. Driehuis, F., Hoolwerf, J., & Rademaker, J. L. W. (2016). Concurrence of spores of Clostridium tyrobutyrium, Clostridium beijerinckii and Paenibacillus polymyxa in silage, dairy cow faeces and raw milk. International Dairy Journal, 63, 70-77. Duanis-Assaf, D., Steinberg, D., Chai, Y., & Shemesh, M. (2016). The LuxS based quorum sensing governs lactose induced biofilm formation by Bacillus subtilis. Frontiers in Microbiology, 6, 1-10. Faille, C., Bénézech, T., Blel, W., Ronse, A., Clarisse, M., & Slomianny, C. (2013). Role of mechanical vs. chemical action in the removal of adherent Bacillus spores during CIP procedures. Food Microbiology, 33, 149-157. Faille, C., Bénézech, T., Midelet-Bourdin, G., Lequette, Y., Clarisse, M., Ronse, G., Ronse, A., & Slomianny, C. (2014). Sporulation of Bacillus spp. within biofilm: A potential source of contamination in food processing environments. Food Microbiology, 40, 64-74. Fernandez-No, I. C., Guarddon, M., Bohme, K., Cepeda, A., Calo-Mata, P., & Barros-Velazquez, J. (2011). Detection and quantification of spoilage and pathogenic Bacillus cereus, Bacillus subtilis and Bacillus licheniformis by real-time PCR. Food Microbiology, 28, 605-610. Ferrand, J., Hadou, T., Selton-Suty, C., Goehringer, F., Sadoul, N., Alauzet, C., & Lozniewski, A. (2013). Cardiac device-related endocarditis caused by Paenibacillus glucanolyticus. Journal of Clinical Microbiology. 51(10), 3439- 3442.

142

Flemming, H. C., & Wingender, J. (2010). The biofilm matrix. Nature Reviews of Microbiology, 8, 623-633. Flint, S. H., Brooks, J. D., & Bremer, P. J. (1997). The influence of cell surface properties of thermophilic Streptococci on attachment to stainless steel. Journal of Applied Microbiology, 83, 508-517. Flint, S. H. (1998). Formation and control of biofilm of thermo-resistant Streptococci on stainless steel. Massey University. New Zealand, Thesis. Flint, S. H., van den Elzen, H., Brooks, J. D., & Bremer, P. J. (1999a). Removal and inactivation of thermo-resistant streptococci colonising stainless steel. International Dairy Journal, 9(7), 429-436. Flint, S. H., Ward, L. J. H., & Brook, J. D. (1999b). Streptococcus waius sp. nov., a thermophilic streptococcus from a biofilm. International Journal of Systematic Bacteriology, 49, 759-767. Flint, S. H., Brooks, J. D., & Bremer, P. J. (2000). Properties of the stainless steel substrate influencing the adhesion of thermo-resistant streptococci. Journal of Food Engineering, 43, 235-242. Flint, S. H., Palmer, J., Bloemen, K., Brooks, J., & Crowford, R. (2001). The growth of Bacillus steathermophilus on stainless steel, Journal of Applied Microbiology, 90, 151-157. Flint, S., Palmer, J., Bremer, P., Seale, B., Brooks, J., Lindsay, D., & Burgess, S. (2011a). Biofilm Formation. Encyclopedia of Dairy Sciences (Second Edition), San Diego: Academic Press, 445-450. Flint, S. H., Tang, X., Bennett, R. J., & Brooks, J. D. (2011b). Controlling the microbiology of whey and whey products. Whey: Types, Composition and Health Implications. Book Chapter, 115-135. Flint, S., Gonzaga, Z. J., Good, J. & Palmer, J. (2017). Bacillus thermoamylovorans- A new threat to the dairy - A review. International Dairy Journal, 65, 38-43. Fouladitajar, A., Ashtiani, F. Z., Rezaei, H., Haghmoradi, A., & Kargari, A. (2014). Gas sparging to enhance permeate flux and reduce fouling resistances in cross flow microfiltration. Journal of Industrial and Engineering Chemistry, 20, 624-632. Frank, J. F. (1997). Milk and dairy products. In: Doyle, M. P., Beuchat, L. R., & Montville, T. J., editors. Food Microbiology: Fundamentals and Frontiers. Washington, D.C.: American Society for Microbiology, 100–116. Ghafoor, A., Hay, I. D., & Rehm, B. H. A. (2011). Role of exopolysaccharides in Pseudomonas aeruginosa biofilm formation and architecture. Applied & Environmental Microbiology, 77(15), 5238-5246. Gilmore, A., & Rowe, M. T. (1990). Microorganisms associated with milk. In: Robinson, R. K. (Ed), Dairy Microbiology, Elsevier Applied Science, London, 37-76. Geiger, B., Nguyen, H-M., Wenig, S., Nguyen, H. A., Lorenz, C., Kitti, R., Mathiesen, G., Eijsink, V. G. H., Haltrich, D., & Nguyen, T-H. (2016). From by-product to valuable components: Efficient enzymatic conversion of lactose

143

in whey using β-galactosidase from Streptococcus thermophilus, Biochemical Engineering Journal, 116, 45-53. Hay, I. D., Gatland, K., Campisano, A., Jordens, J. Z., & Rehm, B. H. (2009). Impact of alginate overproduction on attachment and biofilm architecture of a super mucoid Pseudomonas aeruginosa strain. Applied Environmental Microbiology, 75, 6022-6025. Herzberg, M., & Elimelech, M. (2007). Biofouling of reverse osmosis membranes: Role of biofilm-enhanced osmotic pressure. Journal of Membrane Science, 295, 11-20. Holah, J., & Gibson, H. (2000). Food industry biofilms. In: Allison, D. Gilbert, J. P., Lappin-Scott, H. & Wilson, M., (eds) Community Structure and Co-operation in biofilms. Cambridge University Press, Cambridge, 211-235. Hood, S. K., & Zottola, E. A. (1995). Biofilms in food processing. Food Control, 6(1), 9-18. Huang, C-H., Chang, M-T., Huang, L., & Chu, W-S. (2012). Development of a novel PCR assay based on the gyrase B gene for species identification of Bacillus licheniformis. Molecular and Cellular Probes. 26, 215-217. Huck, J. R., Woodcock, N. H., Ralyea, R. D., & Boor, K. J. (2007). Molecular and subtyping and characterization of psychrotolerant -forming bacteria in two New York state fluid milk processing systems. Journal of Food Protection, 70(10), 2354-2364. Husmark, U. (1993). Adhesion mechanisms of bacterial spores to solid surfaces. Chalmers University of Technology, Sweden, Thesis. Ivy, R. A., Ranieri, M. L., Martin, N. H., den Bakker, H. C., Xavier, B. M., Wiedmann, M., & Boor, K. J. (2012). Identification and characterization of psychrotolerant sporeformers associated with fluid milk production and processing. Journal of Applied and Environmental Microbiology, 78(6), 1853- 1864. Juajun, O., Nguyen, T-H., Maischberger, T., Iqbal, S., Haltrich, D., & Yamabhai, M. (2011). Cloning, purification and characterization of β-galactosidase from Bacillus licheniformis DSM 13. Applied Microbiology Biotechnology, 89, 645-654. Kolodkin-Gal, I., Cao, S., Chai, L., Bottcher, T., Kolter, R., Clardy, J., & Losick, R. (2012). A self-produced trigger for biofilm disassembly that targets exopolysaccharide. Cell, 149, 684-692. Korber, D. R., Lawrence, J. R., Sutton, B., & Caldwell, D. E. (1989). Effects of laminar flow velocity on the kinetics of surface recolonization by mot+ and mot- Pseudomonas fluorescens, Microbiology Ecology, 18, 1-9. Kumar, C. G., & Anand, S. K. (1998). Significance of microbial biofilms in food industry: A review. International Journal of Food Microbiology, 42(1-2), 9- 27.

144

Laborie, S., Cabassud, C., Durand-Bourlier, L., & Laine, J. M. (1998). Fouling control by air sparging inside hollow fibre membranes- effects on energy consumption. Desalination, 118, 189-196. Lagrange, V., Whitsett, D., & Burris, C. (2015). Global market for dairy proteins. Journal of Food Science, 80, S1, 16-22. Langeveld, L. P. M., van Montfort-Quasig, R. M. G. E., Weerkamp, A. H., Waalewijn, R. & Ever, J. S. (1995). Adherence, growth and release of bacteria in a tube heat exchanger for milk. Netherlands Milk and Dairy Journal, 49, 207-220. Lehmann, F. L. (1995). Recontamination of industrial ultrafiltration units and pasteurisers by thermoduric-thermophilic bacteria. International Dairy Federation Symposium on Heat Treatments and Alternative Methods, Vienna. Lemon, K. P., Higgins, D. E., & Kolter, R. (2007). Flagellar motility is critical for Listeria monocytogenes biofilm formation. Journal of Bacteriology, 189(12), 4418-4424. Lin, S. C., Minton, M. A., Sharma, M. M., & Georgiou, G. (1994). Structural and immunological characterization of a biosurfactant produced by a Bacillus licheniformis JF-2. Applied Environmental Microbiology, 60, 31-38. Liu, C., Tanaka, H., Zhang, L., Zhang, J., Huang, X., Ma, J., & Matsuzawa, Y. (2012). Fouling and structural changes of Shirasu porous glass (SPG) membrane used in aerobic wastewater treatment process for microbubble aeration. Journal of Membrane Science. 421, 225-231. Logan, N. A. (2011). Bacillus and relatives in foodborne illness. Journal of Applied Microbiology. 112, 417-429. Logan, N. A., & Vos, P. D. (2015). Bacillus. Bergey’s manual of Systematics of Archae and Bacteria, 1-163. Lorentz, R. H., Ártico, S., da Silveira, A. B., Einsfeld, A., & Corção, G. (2006). Evaluation of antimicrobial activity in Paenibacillus spp. strains isolated from natural environment. Letters in Applied Microbiology, 43, 541-547. Luck, P. J., Vardhanabhuti, B., Yong, Y. H., Laundon, T., Barbano, D. M., & Foegeding, E. A. (2013). Comparison of functional properties of 34% and 80% whey protein and milk serum protein concentrates. Journal of Dairy Science, 96(9), 5522-5531. Lücking, G., Stoeckel, M., Atamer, Z., Hinrichs, J. & Ehling-Schulz, M. (2013). Characterization of aerobic spore-forming bacteria associated with industrial dairy processing environments and product spoilage. International Journal of Food Microbiology, 166, 270-279. Macgibbon J, (2014). Whey to go. Ngaio Press, Martinborough, Wellington, New Zealand. Madaeni, S. S., & Mansourpanah, Y. (2004). Chemical cleaning of reverse osmosis membranes fouled by whey. Desalination, 161, 13 – 24.

145

Madslien., E. H., Rønning, H. T., Lindbäck, T., Hassel, B., Andersson, M. A., & Granum, P. E. (2013). Lichenysin is produced by most Bacillus licheniformis strains. Journal of Applied Microbiology, 115, 1068 – 1080. Marchand, S., Block, J. De., De Jonghe, V., Coorevits, A., Heyndrickx, M. & Herman, L. (2012). Biofilm formation in milk production and processing environments; influence on milk quality and safety. Comprehensive Reviews in Food Science and Food Safety, 11(2), 133-147. Masiello, S. N., Martin, N. H., Watters, R. D., Galton, D. M., Schukken, Y. H., Wiedmann, M., & Boor, K. J. (2014). Identification of dairy farm management practices associated with the presence of psychrotolerant sporeformers in bulk tank milk. Journal of Dairy Science, 97(7), 4083-4096. Mielich-Süss, B., & Lopez, D. (2015). Molecular mechanisms involved in Bacillus subtilis biofilm formation. Environmental Microbiology, 17(3), 555-565. Mikkola, R., Kolari, M., Andersson, M. A., Helin, J., & Salkonija-Salonen, M. S. (2000). Toxic lactonic lipopeptide from food poisoning isolates of Bacillus licheniformis. European Journal of Biochemistry, 267(13), 4068-4074. Miller, R. A., Kent, D. J., Watterson, M. J., Boor, K. J., Martin, N. H., & Wiedmann, M., (2015). Spore populations among bulk tank raw milk and dairy powders are significantly different. Journal of Dairy Science, 98(12), 8492-8504. Mohammad, A. W., Ng, Y. C., Lim., Y. P., & Ng, G. H. (2012). Ultrafiltration in Food Processing Industry: Review on application, Membrane fouling and fouling control. Food Bioprocess and Technology, 5, 1143-1156. Mohammadi, T., Madaeni, S. S., & Moghadam, M. K. (2002). Investigation of membrane fouling. Desalination, 153, 155-160. Morr, C. V. & Ha, E. Y. W. (1991). Off-flavours of whey protein concentrates: A literature review. International Dairy Journal. 1, 1-11. Morr, C. V. & Ha, E. Y. W. (1993). Whey protein concentrates and isolates: Processing and functional properties. Critical reviews in Food science and Nutrition, 33(6), 431-476. Murphy, P. M., Lynch, D., & Kelly, P. M., (1999). Growth of thermophilic spore forming bacilli in milk during the manufacture of low heat powders. International Journal of Dairy Technology, 52, 45-50. Nijland, R., Hall, M. J., & Burgess, J. G. (2010). Dispersal of biofilms by secreted, matrix degrading, bacterial DNase. PLos ONE, 5(12). doi: 10.1371/journal.pone.0015668. O'Toole, G., Kaplan, H. B., & Kolter, R. (2000). Biofilm formation as microbial development. Annual Review of Microbiology, 54, 49-79. Oh, S-W., Chen, P-C., & Kang, D-H. (2007). Biofilm formation by Enterobacter sakazakii grown in artificial broth and infant milk formula on plastic surface. Journal of Rapid Methods and Automation in Microbiology, 15, 311-319. Oknin, H., Steinberg, D., & Shemesh, M. (2015). Magnesium ions mitigate biofilm formation of Bacillus species via downregulation of matrix genes expression. Frontiers in Microbiology, 6, 1-7.

146

Okuku, D. O., Onwulata, C., Mukhopadhyay, S., Chau, L., Thomas-Gahring, A., & Tunick, M. H. (2017). Changes in microbial populations of WPC34 and WPC80 whey protein during long term storage. Journal of Food Processing and Preservation, 41, 1-8. Onwulata, C. I., & Huth, P. J. (2008). Whey processing, functionality and health benefits: IFT Press.Wiley_Blackwell. Palmer, J., Flint, S., & Brooks, J. (2007). Bacterial cell attachment, the beginning of a biofilm. Journal of Industrial Microbiology & Biotechnology, 34(9), 577- 588. Palmer, J. S., Flint, S. H., Schmid, J., & Brooks, J. D. (2010). The role of surface charge and hydrophobicity in the attachment of Anoxybacillus flavithermus isolated from milk powder. Journal of Industrial Microbiology & Biotechnology, 37, 1111-1119. Pamp S. J., & Tolker-Nielsen, T. (2007). Multiple roles of biosurfactants in structural biofilm development by Pseudomonas aeruginosa. Journal of Bacteriology, 189, 2531-2539. Parkar, S. G., Flint, S. H., Palmer, J. S., & Brooks, J. D. (2001). Factors influencing attachment of thermophilic bacilli to stainless steel. Journal of Applied Microbiology, 90, 901-908. Parkar, S. G., Flint, S. H., & Brooks, J. D. (2003). Physiology of biofilms of thermophilic bacilli-potential consequences for cleaning. Journal of Industrial Microbiology & Biotechnology, 30(9), 553-560. Pasvolsky, R., Zakin, V., Ostrova, I., & Shemesh, M. (2014). Butyric acid released during milk lipolysis triggers biofilm formation of Bacillus species. International Journal of Food Microbiology, 181, 19-27. Peeters, E., Nelis, H. J., & Coenye, T. (2008). Comparison of multiple methods for quantification of microbial biofilm grown in microtiter plates. Journal of Microbiological Methods, 72(2), 157-165. Phattaranawik, J., & Leiknes, T. (2011). Extractive biofilm membrane bioreactor with energy recovery and new membrane fouling control. Bioresource Technology, 102, 2301-2307. Poulsen, L. V. (1999). Microbial biofilm in Food processing. Lebensm.-Wiss. u.- Technology, 32, 321-326. Raats, D., Offek, M., Minz, D., & Halpern, M. (2011). Molecular analysis of bacterial communities in raw cow milk and the impact of refrigeration on its structure and dynamics. Food Microbiology, 28, 465-471. Ranieri, M. L., Huck, J. R., Sonnen, M., Barbano, D. M., & Boor, K. J. (2009). High temperature short time pasteurization temperatures inversely affect bacterial numbers during refrigerated storage of pasteurized fluid milk. Journal of Dairy Science, 92(10), 4823-4832. Reginensi, S. M., Gonzàlez, M. J., Olivera, J. A., Sosa, M., Juliano, P., & Bermŭdez, J. (2011). RAPD-based screening for spore forming bacterial populations in

147

Uruguayan commercial powdered milk. International Journal of Food Microbiology. 148, 36-41. Rivardo, F., Martinotti, M. G., Turner, R. J., & Ceri, H. (2011). Synergistic effect of lipopeptide biosurfactant with antibiotics against Escherichia coli CFT073 biofilm. International Journal of Antimicrobial Agents. 37, 324-331. Ron, E. Z. & Rosenberg, E. (2001). Natural roles of biosurfactants. Environmental Microbiology, 3(4), 229-236. Ronimus, R. S., Parker, L. E., & Morgan, H.W. (1997). The utilization of RAPD- PCR for identifying thermophilic and mesophilic Bacillus species. FEMS Microbiology Letters, 147, 75-79. Roux, D., Cywes-Bentey, C., Zhang, Y-F., Pons, S., Konkol, M., Kearns, D., Little, D. J., Howell, P. L., Skurnik, D., & Pier, G. B. (2015). Poly-N- acetylglucosamine as a major polysaccharide component of the B. subtilis biofilm matrix. The Journal of Biological Chemistry, 290, 19261- 19272. Rucket, A., Ronimus, R. S., & Morgan, H. W. (2004). A RAPD-based survey of thermophilic Bacillli in milk powders from different countries. International Journal of Food Microbiology. 96, 263-272. Sadiq, F. A., Li, Y., Liu, T., Flint, S., Zhang, G., & He, G. (2016a). A RAPD based study revealing a previously unreported wide range of mesophilic and thermophilic spore formers associated with milk powders in China. International Journal of Food Microbiology, 217, 200–208. Sadiq, F. A, Li, Y, Liu, T. J., Flint, S., Zhang, G., Yuan, L., Pei, Z., & He, G. (2016b). The heat resistance and spoilage potential of aerobic mesophilic and thermophilic spore forming bacteria isolated from Chinese milk powders. International Journal of Food Microbiology, 238, 193-201. Sayem, S. M. A, Manzo, E., Ciavatta, L., Tramice, C., Cordone, A., Zanfardina, A., Felice, M. D., & Varcamonti, M. (2011). Anti-biofilm activity of an exopolysaccharide from a sponge-associated strain of Bacillus licheniformis. Microbial Cell Factories, 10 (74), 1-12. Scheldeman, P. & Goossens, K. (2004). Paenibacillus lactis sp. nov., isolated from raw and heat-treated milk. International Journal of Systematic and Evolutionary Microbiology, 54, 885-891. Schreiber, R. (2001). Heat-induced modifications in casein dispersions affecting their rennet ability. International Dairy Journal. 11, 553-558. Scott, S. A., Brooks, J. D., Rakonjac, J., Walker, K. M. R., & Flint, S. (2007). The formation of thermophilic spores during the manufacture of whole milk powder. International Journal of Dairy Technology, 69, 109-117. Seale, R. B., Flint, S. H., McQuillan, A. J., & Bremer, P. J. (2008). Recovery of spores from thermophilic dairy bacilli and effects of their surface characteristics on attachment to different surfaces. Applied and Environmental Microbiology, 74(3), 731-737.

148

Seale, R. B., Bremer, P. J., Flint, S. H., & McQuillan, A. J. (2010). Characterization of spore surfaces from a Geobacillus sp. isolate by pH dependence of surface charge and infrared spectra. Journal of Applied Microbiology, 109, 1339-1348. Seale, B., Burgess, S., Flint, F., Brooks, J., Bremer, P., & Parker, S. (2015). Thermophilic Spore-forming Bacilli in the dairy industry. Chapter 7: Biofilms in the dairy industry. Society of Dairy Technology. Wiley Blackwell, UK, 114. Simões, M., Simões, L. C. & Vieira, M. J. (2010). A review of current and emergent biofilm control strategies. LWT- Food Science and Technology, 43, 573-583. Smithers, G. W. (2008). Whey and whey proteins- From ‘gutter-to-gold’. International Dairy Journal, 18, 695-704. Somerton, B., Flint, S., Palmer, J., Brooks, J., & Lindsay, D. (2013). Preconditioning with cations increases the attachment of Anoxybacillus flavithermus and Geobacillus species to stainless steel. Applied and Environmental Microbiology, 79, 4186-4190. Somerton, B., Lindsay, D., Palmer, J., Brooks, J., & Flint, S. (2015). Changes in sodium, calcium, and magnesium ion concentrations that inhibit Geobacillus biofilms have no effect on Anoxybacillus flavithermus biofilms. Applied and Environmental Microbiology. 81(15), 5115-5122 Spiers, A. J., Bahannon, J., Gehrig, S. M., & Rainey, P. B. (2003). Biofilm formation at the air-liquid interface by the Pseudomonas aeruginosa SBW25 wrinkly spreader requires an acetylated form of cellulose. Molecular Microbiology, 50, 15-27. Stanley, R. N., Britton, R. A., Grossman, A. D., & Lazazzera, B. A. (2003). Identification of catabolite repression as a physiological regulator of biofilm formation by Bacillus subtilis by use of DNA microarrays. Journal of Bacteriology, 185(6), 1951-1957. Stewart, P. S., & Costerton, J. W., (2001). Antibiotic resistance of bacteria in biofilms. Lancet, 358, 135-138. Stoeckel, M., Lücking, G., Ehling-Schulz, M., Atamer, Z., & Hinrichs, J. (2016). Bacterial spores isolated from ingredients, intermediate and final products obtained from dairies: thermal resistance in milk. Dairy Science & Technology, 96, 569-577. Tang, X., Flint, S. H., Bennet, R. J., Brooks, J. D., & Morton, R. H. (2009). Biofilm growth of individual and dual strains of Klebsiella oxytoca from the dairy industry on ultrafiltration membranes. Journal of Industrial Microbiology & Biotechnology, 36 (12), 1491 -1497. Tang, X., Flint, S. H., Bennett, R. J., & Brooks, J. D. (2010). The efficacy of different cleaners and sanitisers in cleaning biofilms on UF membranes used in the dairy industry. Journal of Membrane Science, 352, 71-75. Tang, X. (2011). Controlling biofilm development on ultrafiltration and reverse osmosis membranes used in dairy plants. Massey University, New Zealand, Thesis.

149

Tang, X., Flint, S., Bennett, R., Brooks, J., & Zain, S. N. M. (2015). Biofilm in the dairy industry. Chapter 8: Biofilm contamination of ultrafiltration and reverse osmosis plants. John Wiley & Sons. Oxford. Teh, K. H., Flint, S., Palmer, J., Andrewes, P., Bremer, P., & Lindsay, D. (2012). Proteolysis produced within biofilms of bacterial isolates from raw milk tankers. International Journal of Food Microbiology, 157, 28-34. Teh, K. H., Lindsay, D., Palmer, J., Andrewes, P., Bremer, P., & Flint, S. (2013). Lipolysis within single culture and co-culture biofilms of dairy origin. International. Journal of Food Microbiology. 163, 129-135. Teh, K. H., Flint, S., Brooks, J. & Knight, G. (2015). Biofilms in the dairy industry. Society of Dairy Technology, Wiley Blackwell, UK. Van Dyk, J. S., Kee, N. L. A., Frost, C. L., & Pletschke, B. I. (2012). Extracellular polysaccharides production in Bacillus licheniformis SVDI and its immunomodulatory effect. Bioresources, 7(4), 4776-4993. Vlamakis, H., Chai, Y., Beauregard, P., Losick, R., & Kolter, R. (2013). Sticking together: Building biofilm the B. subtilis way. Natural Review of Microbiology, 11(3), 157-168. Vithanage, N. R., Yeager, T. R., Jadhav, S. R., Palombo, E. A., & Datta, N. (2014). Comparison of identification systems for psychrotrophic bacteria isolated from raw bovine milk. International Journal of Food Microbiology, 189, 26-38. Watterson, M. J., Kent, D. J., Boor, K. J., Wiedmann, M., & Martin, N. H. (2014). Evaluation of dairy powder products implicates thermophilic sporeformers as the primary organisms of interest. Journal of Dairy Science, 97, 2487-2497. Wiegand, S., Dietrich, S., Hertel, R., Bongaerts. J., Evers, S., Volland, S., Daniel, R., & Liesegang, H. (2013). RNA-seq on Bacillus licheniformis: active regulatory RNA features expressed within a productive fermentation. BMC Genomics, 14(667), 1- 20. Wiencek, M. K., Klapes, N. A., & Foegeding, P. M. (1990). Hydrophobicity of Bacillus and Clostridium spores. Applied and Environmental Microbiology, 56, 2600-2605. Wiencek, M. K., Klapes, N. A., & Foegeding, P. M. (1991). Adhesion of Bacillus spores to inanimate materials: effects of substratum and spore hydrophobicity. Biofouling, 3, 139-149. Wijman, J. G. E., de Leeuw, P. P. L. A., Moezelaar, R., Zwietering, M. H., & Abee, T. (2007). Air-liquid interface biofilms of Bacillus cereus: Formation, sporulation and dispersion. Applied and Environmental Microbiology, 73(5), 1481-1488. Witthuhn, M., Lücking, G., Atamer, Z., Ehling-Schulz, M., & Hinrichs, J. (2011). Thermal resistance of aerobic spore formers isolated from food products. International Journal of Dairy Technology. 64(4), 486-493. Xiong, Y., & Liu, Y. (2010). Biological control of microbial attachment: a promising alternative for mitigating membrane fouling. Journal of Applied Microbiology & Biotechnology, 86, 825-837.

150

Yakimov, M. M., Timmis, K. N., Wray, V., & Fredrickson, H. L. (1995). Characterization of a new lipopeptide surfactant produced by thermotolerant and halotolerant subsurface Bacillus licheniformis BAS 50. Applied Environmental Microbiology, 61, 1706-1713. Yakimov, M. M., Amro, M. M., Bock, M., Boseker, K., Fredrickson, H. L., & Timmis, K. N. (1997). The potential of Bacillus licheniformis for in situ enhanced oil recovery. Journal of Petroleum Science Engineering, 18, 147- 160. Yang, L., Liu, Y., Wu, H., Song, Z., Hoiby, N., Molin, S., & Givskov, M. (2012). Combating biofilms. Immunology & Medical Microbiology, 65, 146-157. Yoon, J-H., Kim, D. K., Lee, J.-S., Shin, K.-S., Sato, H. H., Lee, S. T., Park, Y. K., & Park, Y.–H. (1998). Paenibacillus campinasensis sp. nov., a cyclodextrin- producing bacterium isolated in Brazil. International Journal of Systematic Bacteriology,48, 833-837. Yuan, D.-D., Liu, G.-C., Ren, G.-Y., Zhang, D., Zhao, L., Kan, C.-P., & Zhang, L.- B. (2012). A survey on occurrence of thermophilic Bacilli in commercial milk powders in China. Food Control, 25, 752-757. Zadow, J. G. (1992). Whey and lactose processing. England: Elsevier Science Publisher Ltd. Zaky, A., Escobar, I., Motlagh, A. M., & Gruden, C. (2012). Determining the influence of active cells and conditioning layer on early stage biofilm formation using cellulose acetate ultrafiltration membrane. Desalination, 286, 296-303. Zain, S. N. M., Flint, S. H., Bennett, R. J., & Soon, T. H. (2016). Characterisation and biofilm screening of the predominant bacteria isolated from whey protein concentrate 80. Dairy Science & Technology, 96, 285. Zain, S. N. M., Bennett, R. J., & Flint, S. H. (2017). The potential source of B. licheniformis contamination during whey protein concentrate 80 manufacture. Journal of Food Science. 82(3), 751-756. doi: 10.1111/1750-3841.13633 Zhao, Y., Caspers, M. P., Metselaar, K. J., de Boer, P., Roeselers, G., Moezelaar, R., Groot, M. N., Montijn, R. C., Abee, T., & Kort, R., (2013). Abiotic and microbiotic factors controlling biofilm formation by thermophilic sporeformers. Applied and Environmental Microbiology, 79, 5652-5660.

151

APPENDICES

152

APPENDIX 1

153

APPENDIX 2

Colony formation by one pure single isolate suspected to be B. cereus on Mannitol yolk polymyxin agar after 24 h incubation at 30°C.

154

APPENDIX 3

1Kb Ladder

1000 800

600 613 bp

400 300 200

100

Positive PCR results for several B. licheniformis isolates

Details Well No. Result Bacteria Isolates Primers M 1 Kb ladder - - 1 D55C01 + B-Lich F-1 2 D55C02 B-Lich R-1 + 3 D55C03 + **targeting 4 D55C04 Gyr B gene + 5 D55C11 + 6 E55C02 + 7 F55C01 + 8 F55C02 + 9 F55C11 - 10 F55C12 +

155

APPENDIX 4

The spore image of P. glucanolyticus was selected as photo of the day in American Society for Microbiology (ASM) Facebook page (7 March 2015)

156

APPENDIX 5

157

158

159

160

161

162